[2000] 1 F.C. 146
T-1698-98
Attorney General of Canada (Applicant)
v.
Public Service Alliance of Canada and Canadian Human Rights Commission (Respondents)
Indexed as: Canada (Attorney General) v. Public Service Alliance of Canada (T.D.)
Trial Division, Evans J.—Ottawa, May 31, June 1, 2, 3, 4, 7, 8, 9, 10; Toronto, October 19, 1999.
Human rights — Judicial review of CHRT decision Treasury Board in breach of CHRA, s. 11 by maintaining differences in wages between male, female employees employed in same establishment performing work of equal value — Methodology adopted by CHRT for selecting male comparators for employees in each level of predominantly female complainants group omitting observations from male population where value of work performed higher/lower than highest/lowest value of work performed by female occupational group — Not error of law — Methodology indirectly comparing wages of employees in complainant group performing work having central tendency of value of that group with wages of employees in predominantly male occupational groups performing work of equal value — S. 11 providing only broad, legal framework within which problems of wage discrimination between men, women to be tackled in light of facts of particular employment situation, evidence of expert witnesses, underlying purposes of statute — Methodology calculated to determine extent of systemic discrimination resulting from application over time of wage policies, practices tending either to ignore, undervalue work typically performed by women, by comprehensively viewing pay practices, policies of employer as affecting wages of men, women — Rational basis in evidence supporting CHRT’s exercise of discretion — (2) CHRT holding not necessary to prove differences in wages paid to men, women performing work of equal value based on sex, once established difference in wages paid to men, women performing work of equal value — CHRA, s. 27(2) authorizing CHRC to issue guidelines setting out extent to, manner in which Act applies — Inference Parliament contemplating CHRC’s acquired expertise more important than political accountability for ensuring appropriate exercise of legislative power — Equal Wages Guidelines, 1986 s. 14 deeming as one groups with which complainant group alleging difference in wages, neither incompatible with terms of grant of statutory power, in light of purposes of Act, nor unreasonable exercise of CHRT’s discretion — (3) CHRT holding occupational groups used only to identify in context of group complaints comparators of opposite gender — References to —occupational group— in Guidelines, ss. 12 to 15 simply referring to groups identified under ss. 12, 13 as predominantly male or predominantly female — S. 15 not mandating comparisons be based on employees in predominantly male occupational groups sampled by group — Even if CHRT committed error of law because s. 15 requiring CHRT to base conclusion on wage curve of predominantly male occupational groups, error not warranting quashing of CHRT’s decision — (4) When wages paid to female employees adjusted upwards pursuant to s. 11 complaint, and in accordance with methodology used, any wage difference thereby created statutorily authorized — Not difference established, maintained by employer — (5) Open to CHRT to adopt annual recalculation method whereby wage gap for each year recalculated by taking into account salary increases paid not only to members of complainant group, but also to employees in predominantly male occupational groups included in segmented line in light of imperfect information available.
Public Service — Judicial review of decision of Canadian Human Right Tribunal (CHRT) Treasury Board in breach of CHRA, s. 11 by maintaining differences in wages between male, female employees employed in same establishment performing work of equal value — CHRT not erring in choice of methodology for selecting male comparators for employees in each level of predominantly female complainants group i.e. omitting observations from male population where value of work performed higher/lower than highest/lowest value of work performed by female occupational group — S. 11 aimed at existence of wage gap disadvantaging women as result of gendered segregation in employment, systemic undervaluation of work typically performed by women — Any wage difference created by s. 11 adjustment statutorily authorized — Would not give rise to s. 11 complaint by males then earning less — Annual recalculation of wage gap reasonable — Benefit to public interest of setting aside decision for error of technical nature outweighed by costs of so doing.
Administrative law — Judicial review — Standard of review — Judicial review of CHRT decision Treasury Board in breach of CHRA, s. 11 by maintaining differences in wages between male, female employees employed in same establishment performing work of equal value — S. 11 not containing objective criteria for determining whether jobs involving different tasks of equal value — Enacted at level of principle — Implementation requiring mastery of range of technical knowledge of considerable sophistication, thorough understanding of workplace — Indicating more than general questions of law, legal reasoning, quasi-constitutional questions involved — Reasonable inference CHRT having more expertise in matter than Court — S.C.C. decisions establishing correctness as standard of review applicable to Tribunal’s interpretation of enabling legislation not determinative — On judicial review within Court’s discretion to grant, refuse relief, even when reviewable error by decision-maker — Judicial review public law proceeding; relief granted to further public interest — Benefit to public interest of setting aside CHRT’s decision for failure to comply with Equal Wages Guidelines, 1986, s. 15 outweighed by costs of so doing — In context, justice further delayed would be justice denied.
Construction of statutes — CHRA, s. 11 providing discriminatory practice for employer to establish, maintain differences in wages between male, female employees employed in same establishment, performing work of equal value — Unlike employment equity legislation addressing underrepresentation of women, minorities in certain employment, s. 11 addressing systemic wage discrimination attributable to historic pattern of job segregation — Within CHRT’s mandate when dealing with complaints under s. 11 to take into account existence of underrepresentation of women in higher paying positions — Parliament aware s. 11 represented more statement of principle than complete prescription — Consistent with Parliament’s intention “living tree” of Act should be nourished by experience of other jurisdictions in dealing with social injustice at which s. 11 aimed — CHRT entitled to rely on evidence of expert witnesses who drew on experience with specialized pay equity legislation — Differences between s. 11, other statutes not so significant as to make more modern pay equity legislation irrelevant to resolution of issues before Tribunal.
This was an application for judicial review of a decision of a Canadian Human Rights Tribunal (CHRT) that Treasury Board, the Department responsible for the federal government’s relations with its employees, was in breach of section 11 of the Canadian Human Rights Act by maintaining “differences in wages between male and female employees employed in the same establishment who are performing work of equal value.” Section 11 provides that it is a discriminatory practice for an employer to establish or maintain differences in wages between male and female employees employed in the same establishment who are performing work of equal value.
In February 1990 the Public Service Alliance of Canada (PSAC) filed a complaint with the Commission on behalf of the six predominantly female groups that had been included in a joint union-management initiative (the JUMI study) examining the issue of the underpayment, throughout the federal public service, of those performing work typically performed by females, alleging that the employer was in breach of section 11 in that employees in the six predominantly female occupational groups were being paid less than employees in the 53 predominantly male groups identified in the study who were performing work of equal value to that of members of the female groups. In January 1991 a three-member Tribunal was appointed to inquire into the complaint. It was agreed that it was not possible to make a direct comparison of the value of the work performed by, and the wages paid to, each individual employee in the complainant occupational groups. When a direct comparison of positions is not possible, subsection 15(1) of the Equal Wages Guidelines, 1986 authorizes the making of an indirect comparison of the value of the work and the wages paid. A commonly used statistical method for making an indirect comparison is through the use of a regression line. This involves locating on a graph, the vertical axis of which represents wages paid in dollars and the horizontal axis of which represents the value of work performed in points, each member of a complainant or comparator group. Lines are drawn through the male-female comparators to allow comparisons to be made. The CHRT adopted the “segmented line” approach for quantifying any wage gap. Observations from the male population where the value of the work performed was higher than that of the highest value of the work performed by members of the female occupational group, and lower than that of the lowest value of the work performed by members of the female occupational group were eliminated. The male comparator was a segment of the composite line and included only the employees in predominantly male occupational groups who were performing with a value that fell within the low and high extremes of the values of the work performed by members of the female group. Treasury Board objected that the two populations were not performing work of equal value merely because the work that they performed had the same lowest and highest values; a determination of whether the groups were performing work of equal value required consideration of the distribution of values between these two points. It argued that the methodology selected for measuring a wage differential between males and females must ensure a similarity between the male comparators as a group, and the female occupational group.
Section 14 of the Equal Wages Guidelines, 1986 provides that when a complainant occupational group alleges a difference in wages with more than one other occupational group, those groups are deemed to be one. The Attorney General argued that section 14 is ultra vires because its effect would be to measure the wage differential by taking into account differences not based on sex. CHRT held that once PSAC had established a difference in the wages paid to men and women performing work of equal value, it had proved its case, subject only to the factors identified in section 16 of the Guidelines that justified differences by reference to various criteria, none of which was applicable. It was not necessary to prove that the differences in wages paid to men and women performing work of equal value were based on sex. The CHRT seems also to have accepted that occupational groups are used only to identify in the context of group complaints comparators of the opposite gender. Finally, the CHRT held that any wage adjustment required was retroactive from the date of the decision (1998) to 1985, and must be recalculated for each year.
The issues were: (1) what was the appropriate standard of review; (2) what principles of statutory interpretation applied; (3) whether the methodology adopted (i) failed to measure differences in wages paid to male and female employees who were performing work of equal value for determination of wage differentials; (ii) did not ensure that the only wage differences measured were those caused by sex discrimination; (iii) failed to compare the wages of the female complainants to those paid to employees in a predominantly male occupational group; (iv) resulted in wage ratcheting; (4) how the amounts payable in each year between 1985 and 1998, except 1987, should be calculated; and (5) whether the Government had so misconducted itself that it ought not to be granted relief.
Held, the application should be dismissed.
(1) The omission of any objective criteria for determining whether jobs that involve very different tasks are of equal value in CHRA, section 11 left the Commission and the Tribunal considerable scope to decide how the principle is to be operationalized in any given employment context. Decisions of the Supreme Court of Canada establishing correctness as the standard of review applicable to the Tribunal’s interpretation of its enabling legislation were not determinative of the issues raised in this case. Section 11 was enacted at the level of a principle, and requires for its implementation mastery of a range of technical knowledge of considerable sophistication, and a thorough understanding of the given workplace. Parliament did not provide a definitional blueprint of such specificity that its implementation in any given context inevitably involves the Tribunal in questions of statutory interpretation, and hence of law, that are reviewable on a standard of correctness in an application for judicial review. That the implementation of a statutory provision calls for a range of technical expertise much broader than that possessed by courts of law is a clear indication that more than general questions of law, legal reasoning or quasi-constitutional values are involved.
Although members of human rights tribunals are appointed ad hoc to hear particular cases, and perform only adjudicative functions, it was reasonable to infer from the length and nature of the hearing, volumes of documentary evidence and duration of the CHRT’s immersion in the issues, that the members of the CHRT likely had a better grasp of the problems of operationalizing the principle of pay equity in the federal public service than a judge would acquire in the course of an eight-and-a-half day hearing of an application for judicial review.
(2) According to the Attorney General, the CHRT erred in relying on the evidence of experts whose testimony was shaped by their experience under pay equity statutory regimes which, in some important respects, differed from the CHRA. In contrast to the principle of employment equity which addresses underrepresentation of women and minority groups in certain employment, the CHRT’s concern was merely to tackle the systemic wage differences between men and women performing work of equal value, differences that are attributable in part to historic patterns of job segregation. It is within the mandate of the Commission and the CHRT when dealing with a complaint under section 11 to take into account the existence of the underrepresentation of women in higher-paying positions. It also seems consistent with the quasi-constitutional nature of the Act that, when it enacts a principle that is not self-applying, it is appropriate to approach its interpretation in light of subsequent experience in dealing with the implementation of that principle, in the absence of any clear indication in the Act that the meaning of section 11 was fixed once and for all at the time of its enactment. Parliament was aware that section 11 represented more a statement of principle than a complete prescription. It is consistent with Parliament’s intention that the “living tree” of the Act should be nourished by the experience of other jurisdictions in dealing with the social injustice at which section 11 is aimed: systemic wage discrimination for work of equal value resulting from the historical segregation of the labour world by gender, and the undervaluation of women’s work. Nothing in the statement of legislative purpose in CHRA, section 2 precludes this approach. Therefore it was not an error of law for the Tribunal to rely on the evidence of expert witnesses who drew on their experience with specialized pay equity legislation.
(3)(i) The CHRT did not err in law when it adopted the segmented line methodology for determining wage differentials. First, the CHRT’s chosen methodology compares, albeit indirectly, through the use of a regression line, the wages of employees in the complainant group performing work that has the central tendency of the value of that group with the wages of employees in predominantly male occupational groups who are performing work of equal value. This is in line with section 11’s proscription of wage differences between men and women performing work of equal value. The Tribunal did not compare the wages paid to employees in the female group performing work at the middle point (or central tendency) of the value of the work performed by members of that group with the wages paid to those performing work at the middle point of the value of the work performed by the comparator group. Second, section 11 provides only a broad legal framework within which problems of wage discrimination between men and women are to be approached in light of the facts of the particular employment situation, the evidence of expert witnesses, and the underlying purposes of the statute. It would be inconsistent with both the underlying purpose of section 11 and the legislative record, to interpret section 11 as impliedly prescribing with particularity the characteristics of the permitted comparative methodologies. Much must be left to be decided by the Commission and the CHRT case by case, with the assistance of experts. Third, the policy motivating the enactment of the principle of equal pay for work of equal value is the elimination from the workplace of sex-based wage discrimination. The discrimination at issue was systemic: it was the result of the application over time of wage policies and practices that tended either to ignore, or to undervalue work typically performed by women. In order to understand the extent of such discrimination in a particular employment context it is important to be able to view as comprehensively as possible the pay practices and policies of the employer as they affect the wages of men and women. The CHRT’s methodology was calculated to achieve this objective because it included data on all employees in predominantly male occupational groups who were performing work that had a value that was not less than, nor more than, the range of the value of the work performed by members of the complainant group. The methodology proposed by the Treasury Board would enable the CHRT to examine only a small amount of the available data. Extending the wage line to include employees in male occupational groups performing work of a higher average value than employees in the complainant group will not necessarily widen the differential measured at the point of average value on the female regression line. There was expert evidence before the CHRT that the wage gap between men and women performing work of equal value tended to widen as the value of the work increased. Given the broad and liberal approach to be taken to the statutory mandate to eliminate systemic wage differentials between male and female employees performing work of equal value, it would not be an error of law for the CHRT to construct a male comparator line, the average value of whose work exceeded the average value of the work performed by the complainant group. Fourth, to the extent that the CHRT’s selection of the segmented line methodology involved the exercise of discretion, it clearly had a rational basis in the evidence. Expert witnesses testified that it was statistically sound, and that it was always preferable to adopt a methodology that incorporated more rather than less data. The CHRT’s methodology was consistent with the methodologies used in other pay equity exercises in jurisdictions where specialized pay equity legislation was in force. Nothing in the Act, or in its underlying policy, lead to the conclusion that the Tribunal’s chosen methodology was inconsistent with the statute.
(ii) Equal Wages Guidelines, 1986, were issued pursuant to subsection 27(2) of the Act, which authorizes the Commission to issue a guideline setting out the extent to which and the manner in which, in the opinion of the Commission, any provision of this Act applies in a class of cases described in the guideline. The provision in question is section 11, which, from its terms and nature inevitably leaves considerable latitude in its implementation. It was inferred from this broad delegation of authority to make subordinate legislation on substantive matters that Parliament must have contemplated that the expertise that the Commission will have acquired in the discharge of its statutory responsibilities for human rights research and public education, and for processing complaints up to the point of adjudication, is more important than political accountability for ensuring the appropriate exercise of the legislative power conferred by subsection 27(2). In view of the breadth of the statutory language of subsection 27(2), and of the attributes of the body to which the discretion has been conferred, a provision in any guidelines issued will only be invalid if it is clearly incompatible with the terms of the grant of statutory power when construed in the light of the purposes of the Act, or is an unreasonable exercise of the Commission’s discretion. Section 14 of the Guidelines is neither.
Section 14 will only be invalid on the first ground if the purpose of the CHRA as set out in section 2 requires words to be read into section 11. Subsection 11(1) means what it says and does not require words to be read into it. First, sections 5 to 10 of the Act, which define other discriminatory practices, include as part of the definition that the proscribed conduct must be “based on a prohibited ground of discrimination”. In contrast, subsection 11(1) does not state that the conduct prohibited must be based on sex. If the reason for that omission was because such an addition would be redundant, it would be unnecessary to include in the definition of relevant conduct proscribed that it was based on one of the prohibited grounds. Second, subsection 11(4) provides that, notwithstanding subsection 11(1), it is not a discriminatory practice to pay different wages to male and female employees performing work of equal value if that difference can be attributed to one of the factors prescribed in guidelines as a reasonable factor justifying the difference. It can be inferred from subsection 11(4) that the definition of the discriminatory practice described in subsection 11(1) is not limited to differences in wages that are based on sex. Otherwise subsection 11(4) would not have been necessary. That it was necessary to exempt from the reach of section 11 wage differences between men and women performing work of equal value that are attributable to prescribed “reasonable factors”, other than sex (subsection 11(5)) is a clear indication that they would otherwise have fallen within the scope of that section. Accordingly, once a complainant has established a difference in the wages paid to male and female employees performing work of equal value, a breach of section 11 is thereby established, subject only to the employer demonstrating that the difference is attributable to one of the “reasonable factors” prescribed in section 16 of the Guidelines. Third, systemic discrimination “results from the simple operation of established procedures of recruitment, hiring and promotion, none of which is necessarily designed to promote discrimination”. Subsection 11(1) tackled the problem of proof by enacting a presumption that, when men and women are paid different wages for work of equal value, that difference is based on sex, unless it can be attributed to a factor identified by the Commission in a guideline as constituting a reasonable justification for it.
The Attorney General did not contend that section 14 was invalid on the ground of unreasonableness.
(iii) The question was whether the Guidelines prescribed the adoption of employer-designated occupational groups as the measuring units for comparing wage differences under section 11. The words “occupational group” recur throughout sections 12 to 15 of the Guidelines, which stipulate some aspects of the manner in which section 11 applies to complaints by groups. Although “occupational group” is not defined in the Guidelines, it was assumed that they have the same meaning in each of these interlinked sections. The references to “occupational group” in these sections of the Guidelines simply refer to the groups identified under sections 12 and 13 as predominantly male or predominantly female. Section 15 does not mandate that comparisons be based on employees in predominantly male occupational groups, sampled by group. Even if the CHRT did commit an error of law because section 15 of the Guidelines requires it to base its conclusion on the wage curve of the predominantly male occupational groups, the error would not warrant the quashing of the Tribunal’s decision.
(iv) The mischief at which section 11 is principally aimed is the existence of a wage gap that disadvantages women, as a result of gendered segregation in employment and the systemic undervaluation of the work typically performed by women. Moreover, section 14 of the Guidelines is an enactment of the Commission’s view that the principle of equal pay for work of equal value does not require the adjustment of female wages up to the level of the highest paid males performing work of equal value. It is sufficient if female wages do not fall below the average of those received by male employees. Accordingly, a complaint by some of the male employees that they are paid less than females whose wages have been adjusted up to the average male wage pursuant to section 11, and that they are thus the victims of a difference in wages contrary to section 11, could not be upheld consistently with the statutory scheme. A literal interpretation of the words of section 11 must yield to an interpretation that is informed by the statutory context in which they appear. When the wages paid to female employees are adjusted upwards pursuant to a section 11 complaint, and in accordance with the methodology used by the CHRT, any wage difference thereby created between the wages of females and males whose wages are below the average could appropriately be regarded as statutorily authorized. It is accordingly not a difference that the employer has established or maintained within the meaning of subsection 11(1). It could also be said that, when a section 11 wage adjustment is made in response to a group complaint, any sex-based wage lines disappear. The result is to produce a single line for all groups regardless of whether they are predominantly male or predominantly female. In such a situation, when a systemic remedy has been fashioned, there is no room for an individual complaint under section 11.
(4) The CHRT’s methodology involved a recalculation of the wage gap for each year by taking into account salary increases paid not only to members of the complainant group but also to employees in the predominantly male occupational group included in the segmented line. The Attorney General objected that this method assumes that the value of the work performed by those whose wages are being compared remained constant. Therefore the wage gap should be calculated in 1987, and increased by a percentage increase in the salaries paid to members of the complainant group. This method does not measure differences in wages between male and female employees performing work of equal value. While wage rate data were available, it would be impossible to determine the actual value of the work performed by members of the complainant and comparator group in the years after 1987. It was open to the Tribunal in these circumstances to adopt the annual recalculation method whereby the wage gap for each year is recalculated by taking into account salary increases paid not only to members of the complainant group, but also to employees in the predominantly male occupational groups included in the segmented line, for determining the wage gap in each of the years covered by the retroactive award period. That Treasury Board’s proposed method of calculating the wage gap proceeds on the equally unlikely assumption that the wage gap remained constant during that period, was taken into account. In light of the imperfect information available, including an absence of any evidence adduced by Treasury Board that changes in the value of work performed by the employees in question had occurred, the CHRT’s methodology was not unreasonable.
(5) On an application for judicial review it is within the Court’s discretion to grant or to refuse the relief sought by an applicant, even when a reviewable error has been committed by the administrative decision-maker. The most powerful reason for withholding relief is that, to set aside the CHRT’s decision on the ground that it did not comply with the Guidelines, section 15 would likely frustrate the purposes underlying CHRA, section 11 and inflict injustice on public servants who would be deprived of backpay. Weighed against these considerations is the principle that public money should not be disbursed pursuant to decisions that are inconsistent with the legally binding instructions of the legislature. An application for judicial review is a public law proceeding and in the final analysis relief is granted by the court in order to further the public interest. Thus relief should be refused when it would not serve the public interest to set aside a decision, even if vitiated by legal error. The following were considered in striking a balance between the competing heads of public interest. (i) In view of the substantial lapse of time, it is highly improbable that reliable data could be assembled from which the value of the work performed by these groups could be determined. (ii) Re-sampling the male comparators by occupational group would involve considerable expense and entail further delay in the resolution of this dispute. Justice unduly delayed in this context is justice denied. (iii) If the Tribunal did breach section 15 of the Guidelines, it was an error of a merely technical nature that did not thwart the essential purposes of section 11 or their implementation. Balancing these competing considerations, the benefit to the public interest of setting aside the Tribunal’s decision for failure to comply with section 15 would be outweighed by the costs of so doing.
It was also argued that the Government was estopped from challenging the validity of the CHRT’s decision as a result of a statement made in 1993 by Mr. Chrétien, then Leader of the Opposition, that if a Liberal Government were elected it would implement without further delay any decision of the Tribunal in this dispute. Accountability for such statements is more appropriately enforced through the political process than by the courts.
STATUTES AND REGULATIONS JUDICIALLY CONSIDERED
Canadian Human Rights Act, R.S.C., 1985, c. H-6, ss. 2 (as am. by S.C. 1998, c. 9, s. 9), 5, 6, 7, 8, 9 (as am. by S.C. 1998, c. 9, s. 12), 10 (as am. idem), 11, 27 (as am. idem, s. 20).
Employment Equity Act, S.C. 1995, c. 44.
Equal Wages Guidelines, 1986, SOR/86-1082, ss. 3, 4, 5, 6, 7, 8, 11, 12, 13, 14, 15, 16.
Federal Court Act, R.S.C. 1970 (2nd Supp.), c. 10, ss. 18, 28.
CASES JUDICIALLY CONSIDERED
APPLIED:
Public Service Alliance of Canada v. Canada (Department of National Defence), [1996] 3 F.C. 789 (1996), 27 C.H.R.R. D/488; 199 N.R. 81 (C.A.); Canadian National Railway Co. v. Canada (Canadian Human Rights Commission), [1987] 1 S.C.R. 1114; (1987), 40 D.L.R. (4th) 193; 27 Admin. L.R. 172; 8 C.H.R.R. D/4210; 87 CLLC 17,022; 76 N.R. 161.
DISTINGUISHED:
Canada (Attorney General) v. Mossop, [1993] 1 S.C.R. 554; (1993), 100 D.L.R. (4th) 658; 13 Admin. L.R. (2d) 1; 46 C.C.E.L. 1; 17 C.H.R.R. D/349; 93 CLLC 17,006; 149 N.R. 1; University of British Columbia v. Berg, [1993] 2 S.C.R. 353; (1993), 102 D.L.R. (4th) 665; 79 B.C.L.R. (2d) 273; 13 Admin. L.R. (2d) 141; 26 B.C.A.C. 241; 18 C.H.R.R. D/310; 152 N.R. 99; 44 W.A.C. 241; Gould v. Yukon Order of Pioneers, [1996] 1 S.C.R. 571; (1996), 133 D.L.R. (4th) 449; 18 B.C.L.R. (3d) 1; 37 Admin. L.R. (2d) 1; 72 B.C.A.C. 1; 25 C.H.R.R. D/87; 194 N.R. 81; 119 W.A.C. 1.
CONSIDERED:
Pushpanathan v. Canada (Minister of Citizenship and Immigration), [1998] 1 S.C.R. 982; (1998), 160 D.L.R. (4th) 193; 11 Admin. L.R. (3d) 1; 43 Imm. L.R. (2d) 117; 226 N.R. 201; Ontario Human Rights Commission and O’Malley v. Simpsons-Sears Ltd. et al., [1985] 2 S.C.R. 536; (1985), 52 O.R. (2d) 799; 23 D.L.R. (4th) 321; 17 Admin. L.R. 89; 9 C.C.E.L. 185; 7 C.H.R.R. D/3102; 64 N.R. 161; 12 O.A.C. 241; Syndicat des employés de production du Québec et de l’Acadie v. Canada (Canadian Human Rights Commission), [1989] 2 S.C.R. 879; (1989), 62 D.L.R. (4th) 385; 11 C.H.R.R. D/1; 89 CLLC 17,022; 100 N.R. 241; Bell Canada v. Communications, Energy and Paperworkers Union of Canada, [1999] 1 F.C. 113 (1998), 167 D.L.R. (4th) 432 (C.A.).
REFERRED TO:
P.S.A.C. v. Canada (Treasury Board) (No. 2) (1996), 29 C.H.R.R. D/349; P.S.A.C. v. Canada (Treasury Board) (1991), 14 C.H.R.R. D/341; Canadian Union of Public Employees, Local 963 v. New Brunswick Liquor Corporation, [1979] 2 S.C.R. 227; (1979), 25 N.B.R. (2d) 237; 97 D.L.R. (3d) 417; 51 A.P.R. 237; 79 CLLC 14,209; 26 N.R. 341.
APPLICATION for judicial review of a decision of a Canadian Human Rights Tribunal (P.S.A.C. v. Canada (Treasury Board) (No. 3) (1998), 32 C.H.R.R. D/349; 98 CLLC 230-031) that Treasury Board, the Department responsible for the federal government’s relations with its employees, was in breach of section 11 of the Canadian Human Rights Act by maintaining “differences in wages between male and female employees employed in the same establishment who are performing work of equal value”. Application dismissed.
APPEARANCES:
Sheila R. Block, Dufferin F. Friesen, Q.C., Lubomyr A. Chabursky, Jane S. Bailey for applicant.
Andrew J. Raven and David Yazbeck for respondent Public Service Alliance of Canada.
Rosemary G. Morgan and Marja Bulmer for respondent Canadian Human Rights Commission.
SOLICITORS OF RECORD:
Tory Tory DesLauriers & Binnington, Toronto and the Deputy Attorney General of Canada, for applicant.
Raven, Allen, Cameron & Ballantyne, Ottawa, for respondent Public Service Alliance of Canada.
Canadian Human Rights Commission Legal Department, Ottawa, for respondent Canadian Human Rights Commission.
The following are the reasons for order rendered in English by
Evans J.:
A. INTRODUCTION
[1] In this application for judicial review the Attorney General of Canada seeks an order to set aside a decision of a Canadian Human Rights Tribunal [(1998), 32, C.H.R.R. D/349] that upheld a complaint made to the Canadian Human Rights Commission by the Public Service Alliance of Canada, the union that represents most employees in the federal public service.
[2] The complaint alleged that the Treasury Board, the Department responsible for the federal government’s relations with its employees, was in breach of section 11 of the Canadian Human Rights Act, R.S.C., 1985, c. H-6 by maintaining “differences in wages between male and female employees employed in the same establishment who are performing work of equal value”.
[3] The principal provisions of the order made by the Tribunal are as follows. The existence and size of any wage gap between the employees in female occupational groups on whose behalf the complaint was filed and employees in male occupational groups is to be determined by the methodology proposed by the Commission. Any wage adjustment that is found to be required is retroactive from the date of the Tribunal’s decision, July 29, 1998, to March 1985, and must be recalculated for each year of this period.
[4] The Attorney General submits that the Tribunal’s decision is vitiated by three fundamental legal flaws resulting from a misinterpretation of section 11 of the Act and of the Equal Wages Guidelines, 1986 [SOR/86-1082], subordinate legislation made by the Commission under section 27 [as am. by S.C. 1998, c. 9, s. 20] of the Act. More particularly, the Attorney General maintains that the Tribunal erred in law by failing to:
(a) adopt a statistical methodology that measured differences in the wages paid to male and female employees who were performing work of equal value; (b) measure differences in wages paid to male and female employees that were based on sex; and (c) compare the wages of the female complainants to those paid to employees in a predominantly male occupational group, or occupational groups.
[5] The Alliance and the Commission deny that the Tribunal made any reviewable error of law. Furthermore, the Alliance submitted, even if it did, the Court should refuse in the exercise of its discretion to set aside the Tribunal’s decision, on the ground that the government had so misconducted itself in this matter that it ought not to be granted relief. Not surprisingly, counsel for the Attorney General vigorously denied this allegation.
[6] In a case where statistics and statistical analysis feature so prominently it is perhaps appropriate that numbers also illustrate the complexity and importance of the issues that it raises. The complaints that started the ball rolling were filed with the Commission in 1984, 15 years ago. The Alliance, which was the complainant before the Tribunal in the present case, is representing some 200,000 federal public service employees who have an interest in the outcome. The Tribunal formally started its hearings on September 9, 1991 and rendered its decision nearly seven years later on July 29, 1998. It held more than 250 days of hearings, and heard witnesses testifying for weeks and, in some instances, months on end. The reasons given by the Tribunal for the decision under review are 200 pages long; in addition, the Tribunal rendered two substantial decisions on preliminary issues [P.S.A.C. v. Canada (Treasury Board) (1991), 14 C.H.R.R. D/341; P.S.A.C. v. Canada (Treasury Board) (No. 2) (1996), 29 C.H.R.R. D/349].
[7] The hearing of this application for judicial review was set down for ten days and lasted for 8” days; the application records, compendia and addenda filed were substantial. The length of these reasons inevitably reflects these factors, and the comprehensive submissions of counsel.
[8] The Treasury Board has made an offer to settle which is now worth approximately $1.3 billion, or about one-quarter of the amount that implementation of the Tribunal’s decision is sometimes alleged to cost. While this is, of course, a very big sum, its size is attributable to the large number of people amongst whom it will be divided (some 200,000), and the 14 years of retroactivity of the Tribunal’s award.
[9] If the Tribunal’s decision is upheld it will result in significant lump-sum payments for individual employees, some of whom are now retired. Those employed for most of the period 1985-1998 stand to gain the most. In addition, the decision will give rise to prospective increases in salaries and pensions.
[10] However, it is also important to keep in proportion the total amounts of money involved by considering them from the perspective of individual employees. On the assumption that the retroactive element of the Tribunal’s decision will cost $5 billion, and that the total number of beneficiaries is 200,000, then the average amount payable to individual employees will be a little less than $2,000 for each year of service, which is subject, of course, to income tax and other employee salary deductions.
[11] Despite the amount of money at stake, and the importance of the issues raised, the disposition of this application for judicial review, like all others, will turn on the statutory grounds on which the Court may review decisions of the Tribunal and the standard of review to be applied, the interpretation of the relevant legislation and its application to the facts.
B. FACTUAL BACKGROUND
[12] My account of the very substantial factual background to the Tribunal’s decision is confined to what is necessary for an understanding of the issues raised in this application for judicial review.
[13] In December 1984 a complaint was filed with the Canadian Human Rights Commission by the Clerical and Regulatory (CR) group of employees in the federal public service. They alleged that members of this predominantly female group were being paid less than members of the predominantly male Public Administration (PA) group, although they were performing work of equal value. This was not the first such complaint to have been filed with the Commission by federal public service employees, but it seems to have precipitated a decision by the government to undertake an examination of the issue of the underpayment, throughout the federal public service, of female employees or, more accurately, of those performing work typically performed by females.
[14] Early in 1985 the public service unions were invited to join this project, which came to be called the Joint Union-Management Initiative (the JUMI). The Canadian Human Rights Commission was also asked to participate, which it did, largely by providing advice to the JUMI Committee that drove the exercise. The Commission also agreed to hold in abeyance the complaint filed in 1984, and not to proceed with any other wage discrimination complaints against the Treasury Board that it received subsequently, pending the completion of the JUMI study.
[15] The study identified nine predominantly female occupational groups and 53 predominantly male groups within the federal public service. These groups varied considerably in size, the largest being the Clerical and Regulatory group which comprised some 50,000 members. Members of a group would not necessarily be doing similar work, nor would the work that they did have a similar value. This was particularly true for the CR group, which was divided into seven discrete levels and spanned a very wide range of work of correspondingly different values.
[16] These occupational groupings had existed for many years and, when collective bargaining was introduced into the federal public service in the late 1960s, they formed the bases of the bargaining units. Not surprisingly, over the years the work actually performed by many employees in any given occupational group is likely to have diverged significantly from their job description. When the JUMI was launched there was no single, system-wide standard in place for measuring the value of work of employees in the federal public service.
[17] The most important achievement of the JUMI was its approval and adoption of a plan for evaluating the work performed by employees on a position-by-position basis, using a common measuring standard. A Seattle firm of management consultants, Norman Willis & Associates, was engaged for the purpose of this task. Mr. Willis had substantial experience in the United States and Canada with similar exercises, although he acknowledged that the JUMI study was the largest pay equity study of which he was aware.
[18] The “Willis plan” involved the design and administration of a very lengthy and sophisticated questionnaire which was designed to discover the true nature and scope of the tasks actually performed by employees in their positions. In particular, the questionnaire attempted to elicit information relevant to the criteria contained in subsection 11(2) of the Act, and elaborated in sections 3 to 8 of the Guidelines, for assessing the value of work on the basis of the levels of skill, effort and responsibility involved, and the conditions under which the work was performed.
[19] The questionnaires were administered to approximately 3,200 employees: 1,800 employees in the nine predominantly female occupational groups, and 1,400 employees in the 53 predominantly male groups. The comparator samples from employees in predominantly male occupational groups were drawn randomly by position occupied, and not by occupational group.
[20] This means that the sample of employees in a particular predominantly male occupational group might represent too small a percentage of the employees in that group to provide statistically reliable information about the group as a whole, especially since there were nearly six times as many male as female occupational groups, and yet the total sample of employees doing “male work” was smaller than that of employees doing “female work”.
[21] The JUMI Committee created sub-committees, whose members were responsible for assessing the value of the work performed by employees on the basis of the information provided in the questionnaires. An exercise of this kind inevitably contains a subjective element, despite the relatively objective nature of the statutory criteria for determining the value of work and the technical expertise that has been developed in assessing the value of work, including an understanding of the ways in which work done by women has often been invisible or systemically undervalued.
[22] The job evaluation exercise not surprisingly gave rise to tensions between union and management members in some sub-committees of the JUMI Committee. There was a suspicion on the employer’s side that sub-committees were inflating the value of the work performed by employees in predominantly female occupational groups, and deflating the value of the work performed by employees in predominantly male occupational groups. If true, this would, of course, have had a tendency to establish or widen a wage differential between male and female employees performing work of equal value.
[23] Even though it means jumping ahead in the chronology, it is important to emphasize at this point that the accuracy of the evaluation of work, or of the assignment of “Willis points” as it became known, is not in issue in these proceedings. This is because the matter was the subject of a motion before the Tribunal, which ruled in its “Phase I” decision that the Treasury Board had not established that differences in the evaluations of work made by the consultants and the sub-committees were attributable to gender bias. This decision was not challenged on an application for judicial review.
[24] For reasons that are not material for present purposes the JUMI study came to an end when the public service unions withdrew their co-operation towards the end of 1989. Shortly thereafter, in January 1990, the Treasury Board unilaterally offered equalization payments to members of three female occupational groups on the basis of the JUMI study, which they accepted.
[25] In February 1990 the Alliance filed a complaint with the Commission on behalf of 6 of the predominantly female groups that had been included in the JUMI study. The complaint alleged that the results obtained by the JUMI study established that the employer was in breach of section 11 of the Canadian Human Rights Act in that employees in the 6 predominantly female occupational groups were being paid less than employees in the 53 predominantly male groups included in the study who were performing work of equal value to that of members of the female groups.
[26] The complaint was investigated by the Commission, which by this time had developed a familiarity with the issues as a result of both dealing with the complaints made in 1984 and earlier, and participating in the JUMI study. The Commission retained Mr. Alan Sunter, a former director of Statistics Canada, to assist it. His expertise was particularly helpful in analysing the data generated by the JUMI study, and in devising a statistical methodology for measuring any gender-based wage differentials in the federal public service.
[27] On the basis of Mr. Sunter’s reports, and other material included in the investigator’s report, the Commission decided on October 16, 1990 to request the appointment of a Tribunal to inquire into the 1990 complaint filed by the Alliance.
[28] In January 1991 a three-member Tribunal was appointed. The chair, Ms. Gillis, had at one time practised as a labour lawyer, but has subsequently worked in human resources. Another member, Mr. Fetterly, is also a lawyer; he has chaired administrative tribunals and has acted as an arbitrator. The third member, Ms. Cowan-McGuigan, is a social worker responsible for the development of programs in New Brunswick to assist those dependent on drugs.
[29] The Tribunal heard from an impressive array of expert witnesses: statisticians (Mr. Sunter and Dr. Shillington called by the Commission, and Dr. Swimmer by the Alliance), and pay equity experts (Dr. Armstrong, Dr. Weiner and Mr. Willis called by the Commission, and Mr. Ranger by the Alliance). In addition, Mr. Sadler and Mr. Durber, Commission officials, gave evidence on the Commission’s understanding of the relevant provisions of the legislation, including the Guidelines that it had promulgated, and the Commission’s view of Mr. Sunter’s methodology for measuring the wage differential between female and male employees performing work of equal value. These witnesses were cross-examined extensively by counsel for the Treasury Board, which called no witnesses of its own.
C. THE STATUTORY FRAMEWORK
[30] The following provisions of the Canadian Human Rights Act are relevant to this application for judicial review.
2. The purpose of this Act is to extend the laws in Canada to give effect, within the purview of matters coming within the legislative authority of Parliament, to the principle that all individuals should have an opportunity equal with other individuals to make for themselves the lives that they are able and wish to have and to have their needs accommodated, consistent with their duties and obligations as members of society, without being hindered in or prevented from doing so by discriminatory practices based on race, national or ethnic origin, colour, religion, age, sex, sexual orientation, marital status, family status, disability or conviction for an offence for which a pardon has been granted.
…
11. (1) It is a discriminatory practice for an employer to establish or maintain differences in wages between male and female employees employed in the same establishment who are performing work of equal value.
(2) In assessing the value of work performed by employees employed in the same establishment, the criterion to be applied is the composite of the skill, effort and responsibility required in the performance of the work and the conditions under which the work is performed.
…
(4) Notwithstanding subsection (1), it is not a discriminatory practice to pay to male and female employees different wages if the difference is based on a factor prescribed by guidelines, issued by the Canadian Human Rights Commission pursuant to subsection 27(2), to be a reasonable factor that justifies the difference.
(5) For greater certainty, sex does not constitute a reasonable factor justifying a difference in wages.
(6) An employer shall not reduce wages in order to eliminate a discriminatory practice described in this section.
…
27. …
(2) The Commission may, on application or on its own initiative, by order, issue a guideline setting out the extent to which and the manner in which, in the opinion of the Commission, any provision of this Act applies in a class of cases described in the guideline.
(3) A guideline issued under subsection (2) is, until it is revoked or modified, binding on the Commission and any member or panel assigned under subsection 49(2) with respect to the resolution of a complaint under Part III regarding a case falling within the description contained in the guideline.
[31] The following sections of the Equal Wages Guidelines, 1986 are also relevant.
Complaints by Individuals
11. (1) Where a complaint alleging a difference in wages is filed by or on behalf of an individual who is a member of an identifiable occupational group, the composition of the group according to sex is a factor in determining whether the practice complained of is discriminatory on the ground of sex.
(2) In the case of a complaint by an individual, where at least two other employees of the establishment perform work of equal value, the weighted average wage paid to those employees shall be used to calculate the adjustment to the complainant’s wages.
Complaints by Groups
12. Where a complaint alleging different wages is filed by or on behalf of an identifiable occupational group, the group must be predominantly of one sex and the group to which the comparison is made must be predominantly of the other sex.
13. For the purpose of section 12, an occupational group is composed predominantly of one sex where the number of members of that sex constituted, for the year immediately preceding the day on which the complaint is filed, at least
(a) 70 per cent of the occupational group, if the group has less than 100 members;
(b) 60 per cent of the occupational group, if the group has from 100 to 500 members; and
(c) 55 per cent of the occupational group, if the group has more than 500 members.
14. Where a comparison is made between the occupational group that filed a complaint alleging a difference in wages and other occupational groups, those other groups are deemed to be one group.
15. (1) Where a complaint alleging a difference in wages between an occupational group and any other occupational group is filed and a direct comparison of the value of the work performed and the wages received by employees of the occupational groups cannot be made, for the purposes of section 11 of the Act, the work performed and the wages received by the employees of each occupational group may be compared indirectly.
(2) For the purposes of comparing wages received by employees of the occupational groups referred to in subsection (1), the wage curve of the other occupational group referred to in that subsection shall be used to establish the difference in wages, if any, between the employees of the occupational group on behalf of which the complaint is made and the other occupational group.
D. THE TRIBUNAL’S DECISION
[32] It is difficult to do justice to the Tribunal’s reasons for decision in a short space: they do, after all, run to some 200 pages. However, because the parties had raised before the Tribunal most of the arguments made in the application for judicial review I can focus my overview of the Tribunal’s reasons on its treatment of the principal issues before it, which are also relevant in this proceeding.
1. Equal Pay for Work of Equal Value
(a) The parties’ positions
[33] Each party proposed a different statistical methodology for determining whether employees in the predominantly female occupational groups included in the Alliance’s complaint were being paid less than employees in predominantly male occupational groups performing work of equal value.
[34] All agreed, however, that the first step in any analysis was to identify and distinguish occupational groups that were predominantly male or predominantly female. Section 12 of the Equal Wages Guidelines, 1986 provides that when a complaint is made on behalf of an occupational group, both it and the occupational group to which comparison is made must be predominantly of different sexes. Occupational groups in which neither sex is predominant are not relevant to this exercise and are accordingly to be ignored.
[35] Whether a group is predominantly of one sex or the other is defined by section 13 of the Guidelines. Thus, a small occupational group of less than 100 members must comprise at least 70% of one sex to be “predominantly” of that sex. On the other hand, a larger occupational group of 500 members or more will be “predominantly” male or female if at least 55% of its members are of that sex.
[36] The complainant groups all satisfied the statutory criteria for being predominantly female, and comparison was only made with employees in occupational groups that were predominantly male.
[37] It was also agreed that it was not possible to make a direct comparison of the value of the work performed by, and the wages paid to, each individual employee in the complainant occupational groups. This was because it was highly unlikely that for each employee in the complainant groups there would be an employee in the comparison groups performing work of equal value. When a direct comparison of positions is not possible, subsection 15(1) of the Guidelines authorizes the making of an indirect comparison of the value of the work and the wages paid.
[38] A commonly accepted statistical method for making an indirect comparison is through the use of a regression line. This involves locating on a graph each member of a complainant or comparator group included in the sample (the observations): the vertical axis of the graph represents dollars (wages paid), and the horizontal, “Willis points” (the value of work performed). Lines can then be drawn through the female and the comparator male observations to “summarize” the data, and allow a comparison to be made between male and female employees.
[39] Nor was it contentious that, since employees in the complainant occupational groups performed work of different values, the value of the work performed by members of a group, as measured in Willis points, should be taken as the central tendency of the group. “Central tendency” is a term that attempts accurately to capture the middle of data, including, in this case, the average (or mean) or median values of the work performed by employees in the predominantly female complainant groups. It was not relevant to the methodology adopted by the Tribunal for it to determine the central tendency of the value of the work performed by employees in predominantly male groups, although, as we shall see, the government’s position was that it should have been.
[40] Thus, the starting point for measuring any wage differential between male and female employees performing work of equal value is to determine the wages paid to employees in the predominantly female group for work valued at the central tendency of the value of the work performed by employees in that group.
[41] The heart of the dispute in this case is over the selection of the appropriate male comparators. The Alliance submitted that the male comparators should comprise all the sampled employees in the 53 predominantly male occupational groups, regardless of the value of the work that they performed. Each of these observations would then be located on a graph that plotted wages against value of work, and a regression line would be drawn through them to produce the “best fit” for the data. This is the “composite line”.
[42] Any wage differential between male and female employees performing work of equal value would then be measured by locating the central tendency of the value of the work of the female occupational group, and drawing a vertical line on the graph from that point to the male composite line. This will identify on the male composite line the wages paid to members of predominantly male occupational groups performing work of equal value to that of the central tendency of the value of the work performed by members of the predominantly female occupational group. The wage differential is the difference between the amounts paid to employees in the male occupational groups and employees in the female group who are performing work that has a value equal to the central tendency of the value of the work of the female group.
[43] The methodology proposed by the Commission was essentially similar, except that it omitted observations from the male population where the value of the work performed was higher than that of the highest value of the work performed by members of the female occupational group, and lower than that of the lowest value of the work performed by members of the female occupational group. That is, the male comparator was a segment of the composite line, and included only the employees in predominantly male occupational groups who were performing work with a value that fell within the low and high extremes of the values of the work performed by members of the female group.
[44] The effect of segmenting the male composite line could be the elimination of a large number of employees in the male groups performing work of a value that exceeded the highest value of work performed by employees in the female group. Even if wages can generally be expected to increase with the value of the work performed, it does not necessarily follow that “trimming” the data in this way will alter the slope of the regression line.
[45] If wages increase proportionately with the value of the work performed, the inclusion of employees performing work of a higher value should have no effect on the slope. If the value of the work performed increases proportionately faster than the wages paid, then the regression line will be flattened by including those performing work of a higher value. The inclusion of those performing work with a higher value will only cause the slope of the regression line to tilt up if the ratio of wages to value of work increases in the case of employees performing work of higher value.
[46] The Treasury Board objected to both the composite and the segmented line methodologies on the ground that they failed to measure wage differentials between males and females performing work of equal value, as required by section 11 of the Canadian Human Rights Act. The essential argument was that, even on the “segmented line” methodology proposed by the Commission, the female group would not be doing work of equal value to the male comparator group if the central tendency of the value of the work performed by members of the comparator male groups was substantially higher than the central tendency of the value of the work of the female group.
[47] In other words, the two populations could not be said to be performing work of equal value merely because the work that they performed had the same lowest and highest values. In order to determine whether the groups are performing work of equal value it is also important to take account of the distribution of values between these two points.
[48] Thus, if most members of the female group were performing work towards the low end of the range, while most members of the male group were performing work at the high end, the groups would not be performing work of equal value. The wages paid to those on the male line performing work of equal value to that of the central tendency of the female group are higher than those of the female group because the slope of the regression line plotting wages against value is drawn up by the higher wages paid to employees in the male group who are performing work of higher value than the central tendency of the value of the work performed by employees in the female groups.
[49] Accordingly, the Treasury Board maintained that, in order to comply with section 11 of the Act, any methodology selected for measuring a wage differential between male and female employees performing work of equal value must ensure a similarity between the male comparators as a group, and the female occupational group. In particular, there must be a similarity in the distribution of the values of the work performed within each of the groups. Only then could it be said that members of the male and female groups were performing work of equal value.
[50] The Treasury Board proposed a methodology that would, in its view, satisfy the statutory requirements: it became known as “the whole group methodology”. For present purposes, it is sufficient to note that its essential element is a requirement that a predominantly male occupational group is only a legally valid comparator if the central tendency of the value of the work performed by its members is the same as, or very close to, that of the female occupational group.
[51] The “whole group” methodology would reduce the amount of any wage differential between male and female employees performing work of equal value if the wages paid to employees performing work of a higher value increase more rapidly in relation to the value of the work performed than is the case for those performing work that has less value.
(b) The decision
[52] The Tribunal decided to adopt the methodology proposed by the Commission for selecting the appropriate male comparators for employees in each level of the predominantly female complainants groups. That is, it used the “level to segment” approach for identifying and quantifying any wage gap. It supported its conclusion in three ways.
[53] First, the purpose of section 11 of the Canadian Human Rights Act is to redress systemic wage discrimination resulting from the historic undervaluation of the work performed by women. It is well established that human rights legislation is quasi-constitutional in nature and that its provisions should be given a broad and liberal interpretation so as to further its underlying purposes.
[54] Second, section 11 focusses on work of equal value performed by male and female employees. Its terms do not indicate that the relevant comparison is to the value of work performed by groups of employees. Admittedly, when a complaint is made on behalf of an occupational group, the Equal Wages Guidelines, 1986 require the consideration of groups. However, this is for the limited purpose of determining whether the complainant groups are predominantly male or female, and ensuring that comparator groups are predominantly of the other sex.
[55] Third, the expert witnesses who testified before the Tribunal all favoured the composite line methodology, whether segmented or not, over the Treasury Board’s “whole group” approach. The evidence of the pay equity experts was that this was the way in which pay equity exercises were commonly approached under provincial pay equity legislation, and in those states in the United States where similar legislation is in effect.
[56] The statisticians testified that the composite or segmented line methodologies were preferable because they incorporated the greatest number of the observations collected from the questionnaires administered to employees working in predominantly male occupational groups. In contrast, the Treasury Board’s methodology would require jettisoning large quantities of data because they did not relate to those employed in occupational groups where the central tendency of the value of the employees’ work approximated that of the female complainant occupational groups.
[57] Moreover, since the populations of employees in male and female groups had been sampled on the assumption that the comparison would be by way of a composite line, not particular occupational groups, there might well be insufficient data to enable a statistically reliable comparison to be made using the Treasury Board’s methodology.
2. Wage Differentials, Sex and Causation
(a) The parties’ positions
[58] The Treasury Board further supported its statistical methodology by arguing that it, or something like it, was legally required because the Tribunal’s jurisdiction was limited to redressing wage differences that were based on sex. Wage differences attributable to other factors, such as bargaining power or market forces, must therefore be disregarded.
[59] In order to ensure that it was redressing only wage differences caused by systemic sex discrimination the Tribunal was required to compare the wages paid to employees in the complainant group for the work performed at the central tendency of the value of the group’s work to wages paid for work of the same value in a predominantly male occupational group with a similar central tendency, provided that the group was not paid more than other male groups performing work of equal value. Since wage differences between two predominantly male occupational groups cannot be based on sex, the wages of members of a more highly paid male occupational group must therefore be ignored for the purpose of section 11.
[60] The Treasury Board further submitted that, to the extent that section 14 of the Guidelines was inconsistent with this position, it was ultra vires the Act. This section provides that, when a complainant occupational group alleges a difference in wages with more than one other occupational group, those groups are deemed to be one. Thus, the effect of section 14, if valid, would be to measure the wage differential by taking into consideration differences that cannot be said to be based on sex.
[61] The Alliance’s position was that once it had established a difference in the wages paid to men and women performing work of equal value it had proved its case, subject only to the factors identified in section 16 of the Guidelines that justified differences by reference to various criteria, none of which was applicable in this case.
[62] Given the difficulties of establishing causation in this area it would be contrary to a broad and purposive interpretation of the Act, counsel argued, to require proof that differences in wages paid to men and women performing work of equal value were based on sex. Accordingly, the validity of section 14 is unimpeachable.
(b) The decision
[63] The Tribunal adopted the position advanced on behalf of the Alliance, which the Commission also supported, for essentially the reasons outlined in the last two paragraphs. By calculating the wage differential for the purpose of section 11 on the basis of the wages paid to men and women for work of equal value (that is, the central tendency of the value of the work of the predominantly female group), the Tribunal accomplished what it termed “on-average fairness”.
[64] The Tribunal meant by this that it did not measure the wage differential by reference to either the highest paid or the lowest paid employees in a predominantly male occupational group performing work of equal value to that of the central tendency of the female group. By treating all employees in the segmented line as a single occupational group, the Tribunal was content to countenance the fact that some men would still be paid more than women for work of equal value, while other men might be paid less.
[65] Accordingly, the attack on the validity of section 14 of the Guidelines, which envisages this result, failed. Again, the Tribunal found support for its conclusion in the testimony of the expert witnesses.
3. Wage Differentials and Occupational Groups
(a) The parties’ positions
[66] The third principal submission advanced by the Treasury Board against the methodologies proposed by both the Alliance and the Commission was that they were incompatible with the Equal Wages Guidelines, 1986. In particular, the Treasury Board argued that section 15 of the Guidelines unmistakably requires that, when a complaint alleges differences in wages contrary to the Act between an occupational group and any other occupational group, the matter is to be determined on the basis of a comparison of employees in those occupational groups.
[67] The complaint filed by the Alliance in 1990 fell within section 15. Accordingly, any methodology used by the Tribunal must be based on a comparison of occupational groups. It could not be said that either all the employees sampled in predominantly male occupational groups, or those in the segmented composite line, constituted an occupational group for the purpose of section 15.
[68] The Alliance submitted that it would be contrary to the statutory scheme to restrict the male comparators in this way. Given that occupational groups are unilaterally determined by the employer and that, with the changing nature of work, they will often have lost whatever coherence they once may have had, their importance in implementing statutory pay equity provisions should be limited. Accordingly, the legal significance of occupational groups in the Guidelines should be restricted to a determination of the predominant sex in any occupational group so that the relevant populations for comparison purposes can be identified.
(b) The decision
[69] It is not clear precisely how the Tribunal dealt with this issue. In its discussion of section 15 of the Guidelines, the Tribunal noted that the section provided for an indirect comparison of wages paid for work of equal value when a complaint is made by an occupational group and a direct comparison, job by job, is not possible. The Tribunal seems also to have accepted the Alliance’s submission that a purposive interpretation of the Guidelines leads to the conclusion that occupational groups play only a limited role in the implementation of the statutory principle of equal pay for work of equal value. That is, occupational groups are used only to identify in the context of group complaints comparators of the opposite gender.
[70] In any event, since the Tribunal adopted a methodology that the respondent had submitted was inconsistent with section 15 of the Guidelines it must have rejected the Treasury Board’s interpretation of that provision.
E. ISSUES AND ANALYSIS
[71] In short, counsel for the Attorney General submitted that the Tribunal had misinterpreted section 11 of the Act and the Guidelines, and had consequently erred in law in adopting the Commission’s methodology for determining and measuring any wage difference between employees in the complainant groups, or in levels within those groups, and employees in predominantly male occupational groups.
[72] I shall deal in turn with each of the issues raised by the Attorney General but, first, it is important to deal with two issues that will help to frame the analysis.
1. The Standard of Review
[73] The standard of review to be applied to the Tribunal’s interpretation of its enabling legislation is not open for debate in this Court. The Supreme Court of Canada has firmly established that the doctrine of curial deference does not apply to tribunals’ interpretation of human rights legislation, including the Canadian Human Rights Act, and that it is the function of the reviewing court to interpret for itself the provisions of the Act that are in dispute: Canada (Attorney General) v. Mossop, [1993] 1 S.C.R. 554. This means that, if the reviewing court’s interpretation differs from that of the Tribunal, then the Tribunal has erred in law and its decision is liable to be set aside.
[74] However, the scope of these propositions and their application to the facts of this case require some discussion.
(a) Statutory interpretation
[75] First, it is important to emphasize that section 11 of the Canadian Human Rights Act is the only provision in the Act that addresses differences in wages paid to male and female employees performing work of equal value, and that only subsection 11(1) is directly relevant to the issues in dispute in this litigation.
11. (1) It is a discriminatory practice for an employer to establish or maintain differences in wages between male and female employees employed in the same establishment who are performing work of equal value.
[76] While this subsection is undoubtedly elegant in its brevity, the absence of the more detailed elaboration typically found in pay equity legislation in the provinces, and in those American states that have adopted it, inevitably leaves considerable scope to the Commission and the Tribunal, with the assistance of pay equity and compensation experts, and statisticians, to decide how the principle is to be operationalized in any given employment context.
[77] Indeed, Parliament was aware that answers to many questions about the implementation of equal pay for work of equal value were not to be found in the lapidary language of section 11. Thus, in introducing the provisions, the Minister of Justice and Attorney General of the day, the Hon. S. R. Basford, responded to an observation from a member to the effect that there were no objective criteria for determining whether jobs that involve very different tasks were nonetheless of equal value.
[78] Noting that some provincial legislatures had been deterred by the difficulties of going beyond proscribing sex-based wage differentials for equal work, he said:
The federal government has adopted a different approach: that we should legislate the principle [of equal pay for work of equal value] and, through the Commission, and through its efforts at setting out guidelines, solve these problems. (Respondents’ Joint Record, vol. IV, at pp. 1138-1139.)
I have no doubt that Mr. Basford would have expressed a similar sentiment with respect to the methodological problems with which the Commission and the Tribunal have struggled in this case.
[79] In short, the correct interpretation of section 11 in my opinion is that Parliament intended to confer on the agencies created to administer the Act a margin of appreciation in determining on a case-by-case basis, and with the assistance of the technical expertise available, how the statutorily endorsed principle of equal pay for work of equal value is to be given effect in any given employment setting.
[80] In my view, decisions of the Supreme Court of Canada establishing correctness as the standard of review applicable to the Tribunal’s interpretation of its enabling legislation are not determinative of the issues raised in this case. Section 11 is a statutory provision that was enacted at the level of a principle, and requires for its implementation mastery of a range of technical knowledge of considerable sophistication, and a thorough understanding of the given workplace.
[81] In contrast, the interpretation of the statutory terms “family status” (Mossop, supra); “services or facilities customarily available to the public” (University of British Columbia v. Berg , [1993] 2 S.C.R. 353); and “offering or providing services, goods or facilities to the public” (Gould v. Yukon Order of Pioneers , [1996] 1 S.C.R. 571), could quite properly be said to have involved general questions of law or legal reasoning that depended principally on the interpretative and other legal skills ascribed to judges and on the understanding of equality and discrimination that courts have derived from their experience with constitutional litigation.
[82] It is true that the delineation of the precise scope of the statutory provisions in issue in those cases, when considered in the context of their particular facts, may not have been plain and obvious. Nor, however, were they the enactment of the kind of open-textured principle found in section 11 that requires regular elaboration before its ready application to any given employment context.
[83] Reverting to section 11, I cannot attribute to Parliament an intention that, by enacting the principle of equal pay for work of equal value, it thereby provided a definitional blueprint of such specificity that its implementation in any given context inevitably involves the Tribunal in questions of statutory interpretation, and hence of law, that are reviewable on a standard of correctness in an application for judicial review.
[84] The fact that the implementation of a statutory provision calls for a range of technical expertise much broader than that possessed by courts of law is a clear indication that more than general questions of law, legal reasoning or quasi-constitutional values are involved.
(b) The expertise of the Tribunal
[85] One reason why the Supreme Court of Canada has selected the correctness setting on the standard of review dial for reviewing legal questions decided by human rights tribunals is that members of these tribunals are typically appointed ad hoc to hear particular cases, and perform only adjudicative functions. Thus, they do not have the expertise of members of permanent agencies with regulatory and policy-making responsibilities for the area of activity within their mandate.
[86] These observations are, of course, applicable to the Tribunal members whose decision is under review here. I would note, however, that the Tribunal held over 250 days of hearings, many of which apparently resembled educational seminars conducted by the expert witnesses for the benefit of the parties and the Tribunal, studied volumes of documentary evidence and lived with this case for seven years. It is reasonable to infer from this that the members of the Tribunal were likely to have a better grasp of the problems of operationalizing the principle of pay equity in the federal public service than a judge would probably be able to acquire in the course of even an 8” day hearing of an application for judicial review.
[87] Counsel for the Attorney General submitted that, despite the understanding of the methodological and policy issues that the Tribunal may have acquired, it had no expertise in law. Indeed, she suggested, the Tribunal may well have lost its legal bearings in the course of struggling to get on board the technical expertise of the statisticians and pay equity consultants.
[88] With all respect, this is not the impression that I have from reading the Tribunal’s reasons for decision, which give prominence to the relevant statutory provisions and to the interpretative guidelines contained in the relevant jurisprudence. I would also add that two members of the Tribunal were legally qualified, with professional experience in the law and practice of labour relations and employment.
[89] While I may give no deference to the Tribunal’s views on the interpretation of the legislation this does not mean that I should be unwilling to be educated by their reasons for decision. Nor, in light of the Tribunal’s lengthy immersion in the issues raised by this dispute and the open-ended nature of the relevant statutory standard, should I be alert to brand as a question of law that which is only doubtfully so, to paraphrase the felicitous words of Dickson J. (as he then was) in Canadian Union of Public Employees, Local 963 v. New Brunswick Liquor Corporation, [1979] 2 S.C.R. 227.
[90] A desire not to expand issues of interpretation so that they encroach on functions that fall within the sphere of expertise of human rights tribunals seems to have prompted Bastarache J. in Pushpanathan v. Canada (Minister of Citizenship and Immigration), [1998] 1 S.C.R. 982, at page 1017 to observe that:
… the precise degree of deference which should be accorded to human rights tribunal may still be open to question. …
2. The Interpretation of the Legislation
[91] Counsel were in agreement that a broad and purposive interpretation must be given to section 11 of the Act, and the Guidelines enacted to put some flesh on the otherwise spare statutory skeleton. The Supreme Court of Canada has consistently held that the quasi-constitutional nature of human rights legislation demands no less. Beyond this, however, there were important differences in the ways in which counsel urged me to approach the interpretation of the relevant statutory provisions.
[92] Counsel for the Attorney General emphasized that, while some provincial legislatures have enacted specialized pay equity legislation to deal with wage differences between male and female employees performing work of equal value, the Parliament of Canada has been content to leave the problem within the confines of the Canadian Human Rights Act. She argued that specialized pay equity legislation, enacted in Manitoba, Nova Scotia, New Brunswick, Prince Edward Island, Ontario and Quebec, typically has features that are not found in a more traditional, complaints-driven human rights statute, such as the Act under consideration here. For example, pay equity legislation typically provides for proactive administration; is administered by its own highly specialized agencies; contains detailed and specific provisions on the implementation of pay equity; and is concerned with pay of equal or comparable worth.
[93] Accordingly, section 11 must be read in the context of the Canadian Human Rights Act and, in particular, section 2. This provides that the purpose of the Act is to give effect to the principle that individuals should have an equal opportunity to make for themselves the lives that they are able, consistent with their responsibilities as members of society, “without being hindered in or prevented from doing so by discriminatory practices based on … sex”.
[94] Thus, it was argued, the Act did not address the structural inequalities that arise from the social expectations, assumptions and norms that have resulted in the underrepresentation of women in many of the better remunerated sections of the labour market. This problem is one of employment equity, a concept that is much broader in scope than that at which section 11 is aimed, namely the elimination of sex-based differences in pay for work of equal value performed by men and women. To the limited extent that the Act deals with the denial to women of access to higher paying jobs it does so through sections 7 and 10 [as am. by S.C. 1998, c. 9, s. 13].
[95] It follows, also, from the failure of the Tribunal to appreciate the relatively narrow focus of section 11, by virtue of its inclusion in a complaint-driven and general statutory human rights scheme, that the Tribunal erred when it relied on the evidence of expert witnesses whose testimony was shaped by their experience under pay equity statutory regimes which, as noted above, are different in some important respects from the Canadian Human Rights Act.
[96] I do not find this analysis persuasive. First, the principle of employment equity, as enacted in the Employment Equity Act, S.C. 1995, c. 44, and at one time the law in Ontario, addresses the underrepresentation of women and members of minority groups in particular types of work or places of employment by increasing the access to those positions by members of the excluded group. In contrast, the Tribunal’s concern was merely to tackle the systemic wage differences between men and women performing work of equal value, differences that are attributable in part to historic patterns of job segregation.
[97] In my opinion it is squarely within the mandate of the Commission and the Tribunal when dealing with a complaint under section 11 to take into account the existence of the underrepresentation of women in higher-paying positions. This is implicit in the judgment of Hugessen J.A. in Public Service Alliance of Canada v. Canada (Department of National Defence), [1996] 3 F.C. 789 (C.A.), where he adopted in the context of section 11 the analysis of systemic discrimination, resulting in part from employment segregation, with which Dickson C.J. had held in Canadian National Railway Co. v. Canada (Canadian Human Rights Commission), [1987] 1 S.C.R. 1114, the Act was intended, on a broad and liberal interpretation, to deal.
[98] Second, it seems to me consistent with the quasi-constitutional nature of the Act that, when it enacts a principle that is clearly not self-applying, it is appropriate to approach its interpretation in light of subsequent experience in dealing with the implementation of that principle, in the absence of any clear indication in the Act that the meaning of section 11 was fixed once and for all at the time of its enactment.
[99] As I have already noted, Parliament was aware that section 11 represented more a statement of principle than a complete prescription: the Minister’s acknowledgement that the difficulties of defining the value of work would have to be tackled by the Commission so as to give it concrete expression in a given set of facts is equally applicable to other aspects of the section. The considerations are not present in this case that persuaded the Court in Mossop, supra, to conclude that, in amending the Canadian Human Rights Act, Parliament had fixed the meaning of “family status” at the time of enactment so as to exclude same-sex couples. In particular, the Court emphasized the fact that, when adding “family status” as a prohibited ground of discrimination, Parliament also rejected an amendment that would have added sexual orientation to the list of proscribed grounds.
[100] Accordingly, it is consistent with Parliament’s intention that the “living tree” of the Act should be nourished by the experience of other jurisdictions in dealing with the social injustice at which section 11 is aimed: systemic wage discrimination for work of equal value resulting from the historical segregation of the labour world by gender, and the undervaluation of women’s work. I see nothing in the statement of legislative purpose in section 2 that precludes this approach.
[101] Counsel for the Attorney General repeatedly emphasized that section 11 was not “pay equity legislation”, but human rights legislation. Indeed, in Bell Canada v. Communications, Energy and Paperworkers Union of Canada, [1999] 1 F.C. 113 (C.A.), at page 118, Décary J.A. noted that:
The words “pay equity” have yet to make their way into federal legislation or regulation ….
“To avoid confusion”, he suggested (at page 119), the precise words of section 11 should be used to describe its subject-matter, namely, “wage discrimination complaints”.
[102] On the other hand, Hugessen J.A. in Department of National Defence, supra expressed no such concerns about nomenclature, stating (at page 795) in a case involving section 11 of the Canadian Human Rights Act: “This case concerns pay equity”. Indeed, throughout his reasons for judgment, Hugessen J.A. used the term “pay equity” to describe the nature of the litigation.
[103] For my part, I do not think that much turns on whether section 11 is characterized as a “pay equity” provision, although it is appropriate to note that pay equity is commonly understood to refer to the principle contained in subsection 11(1), namely that there should be no differential in the wages paid to male and female employees of the same establishment performing work of equal value. On the other hand, it should also be remembered that specialized pay equity legislation may specifically prescribe particular approaches to issues and be intended to be used by the parties in a proactive manner as the basis for adjusting wages.
[104] It was also suggested that the fact that most pay equity legislation enacted the principle of equal pay for work of equal or comparable value gives it a broader scope than the Canadian Human Rights Act, where the narrow term “equal value” is used. However, “comparable worth” would seem simply to refer to work that is found to be of equal value as a result of an indirect comparison through the use of wage curves or regression lines. This is, of course, a technique that is expressly permitted by section 15 of the Guidelines.
[105] In my opinion, therefore, it was not an error of law for the Tribunal to rely on the evidence of expert witnesses who drew on their experience with specialized pay equity legislation. The differences between section 11 and these other statutes are not so significant as to make the more modern pay equity legislation irrelevant to the resolution of the issues before the Tribunal.
3. “Work of Equal Value”: the Definitional Problem
(a) The Attorney General’s position
[106] Counsel for the Attorney General reiterated, refined and elaborated the submissions that had been made to the Tribunal on behalf of the Treasury Board. Their essential thrust, it will be recalled, was that the segmented line methodology was contrary to the Act because it involved the comparison of groups of employees who were not performing work of equal value. To the extent that this approach compares the wages of the members of either a female complainant group, or of a level within a particular female group, to the wages of employees in predominantly male occupational groups, the members of which perform work with a significantly higher central tendency value, it could not be said to be measuring wage differences between male and female employees performing work of equal value.
[107] Counsel sought to demonstrate the arbitrariness of the segmented line methodology in various ways. For example, the existence of a single observation, way out at the high end of the female value line, could bring into the segmented male composite line a relatively large number of employees in predominantly male occupational groups performing work of a much higher value than all but one of the observations from the female occupational group. Since employees performing work of higher value are likely to be paid accordingly, their inclusion will tend to draw up the slope of the regression line, and thus widen the wage differential between men and women as measured at the central tendency value of the work performed by employees in the female group.
[108] In addition, the segmented line could include the “tails” of male occupational groups, most of the members of which perform work that is so much higher in value than the work of anyone in the female occupational group that they could not be said to be part of the same employee population. For example, entry-level lawyers in the public service might be performing work that has the same value as that performed by employees at the high end of the value-of-work line of the female occupational group. However, starting salaries for lawyers, and for others in occupations for which there is a strong market demand, are likely to be set in relation to the salaries of more senior members in their occupational groups, who perform work of much higher value than anyone in the female complainant group, and in response to market forces. The inclusion of these “tails” of occupational groups, nearly all of whose members perform work of a much higher value than anyone in the female complainant group, will inflate the wage differential when measured at the central tendency value of the work performed by the female occupational group.
[109] Counsel also used an analogy which, she argued, showed the basic flaw in the position of the Alliance and the Commission. If you wanted to compare the levels of educational attainment of the inhabitants of towns A and B, it would not be sufficient to know that in the last 10 years each had had one drop-out from grade school, and one high school graduate who went on to become a law professor. Any meaningful comparison of the levels of educational attainment of these groups would obviously also depend on the distribution of the populations between these two extremes. Thus, if 10% of the high school graduates from town A proceeded to university, as compared to 50% from town B, you would not say that their levels of educational attainment were equal simply because they had identical lowest and highest achievers.
(b) Analysis
[110] Despite the attractiveness of counsel’s computer-assisted presentations, I am not satisfied that she established that the Tribunal erred in law by misinterpreting section 11 when it adopted the segmented line methodology for determining wage differentials. I do not accept that section 11 prescribes as precisely as counsel contended the characteristics that a methodology must possess for determining the existence of a prohibited wage gap. Nor can it be said that, to the extent that the Tribunal’s selection of a methodology involved the exercise of discretion, it exercised its discretion unreasonably, or without regard to the evidence before it.
[111] First, it is hard to escape the fact that the Tribunal’s chosen methodology compares, albeit indirectly through the use of a regression line, the wages of employees in the complainant group performing work that has the central tendency of the value of that group with the wages of employees in predominantly male occupational groups who are performing work of equal value. This would seem to be exactly in line with section 11’s proscription of wage differences between men and women performing work of equal value.
[112] It is important to emphasize that the Tribunal did not compare the wages paid to employees in the female group performing work at the middle point (or central tendency) of the value of the work performed by members of that group with the wages paid to those performing work at the middle point of the value of the work performed by the comparator group. Hence, counsel’s analogy of how one would measure the levels of educational attainment of two populations did not advance her case.
[113] Of course, the way in which achievement within a population is distributed between the polar extremes will affect its average level of education, and therefore how well educated it may be said to be. In addition, the person with an average level of educational attainment within the better educated population may be expected to be performing more highly valued work and earning a higher wage than the person in the other population who has the average level of education for that group.
[114] However, there is no a priori reason to think that there would be a wage differential in favour of the person in the better educated population who had the same level of education as the person of average education in the less well educated group. There would be a wage differential if the wages of those in the better educated group with high levels of education rose disproportionately quickly as compared to the wages of those of lower educational attainment.
[115] Second, section 11 is silent on how the male comparator group is to be constructed when it is not possible to make a direct comparison of wages paid to male and female employees engaged in work of equal value. In the absence of any statutory instructions on this matter I do not agree that the principle of equal pay for work of equal value, even in the context of the Canadian Human Rights Act as opposed to more modern, specialized pay equity legislation, mandates that comparators be limited to predominantly male occupational groups performing work that has a central tendency value that is the same as, or very similar to, that of the complainant group.
[116] Section 11 provides only a broad legal framework within which problems of wage discrimination between men and women are to be tackled in light of the facts of the particular employment situation, the evidence of expert witnesses, and the underlying purposes of the statute. In my view it would be inconsistent with both the underlying purpose of section 11, and the legislative record, to interpret the section as impliedly prescribing with the particularity suggested by counsel for the Attorney General the characteristics of the permitted comparative methodologies. Much must inevitably be left to be decided by the Commission and the Tribunal case by case, with the assistance of experts.
[117] Third, the policy motivating the enactment of the principle of equal pay for work of equal value is the elimination from the workplace of sex-based wage discrimination. The kind of discrimination at issue here is systemic in nature: that is, it is the result of the application over time of wage policies and practices that have tended either to ignore, or to undervalue work typically performed by women.
[118] In order to understand the extent of such discrimination in a particular employment context it is important to be able to view as comprehensively as possible the pay practices and policies of the employer as they affect the wages of men and women. The Tribunal’s methodology was calculated to achieve this objective because it included data on all employees in predominantly male occupational groups who were performing work that had a value that was not less than, nor more than, the range of the value of the work performed by members of the complainant group. In contrast, the methodology proposed by the Treasury Board would enable the Tribunal to examine only a small amount of the available data.
[119] Extending the wage line to include employees in male occupational groups performing work of a higher average value than employees in the complainant group will not necessarily widen the differential measured at the point of average value on the female regression line.
[120] If the adoption of the segmented line methodology did have this effect it would be because the wages paid to employees in predominantly male occupational groups, who performed work of a higher average value than employees in predominantly female groups, rose proportionately faster than the value of the work that they performed.
[121] Indeed, there was expert evidence before the Tribunal that the wage gap between men and women performing work of equal value tended to widen as the value of the work increased: for example, Dr. Shillington, R.J.R., vol. VII, at pages 2465-2466.
[122] Suppose, then, that this were true for employees in the federal public service, so that the wage gap widened as the value of the work increased, and that women were underrepresented among employees performing more highly valued work. Given the broad and liberal approach to be taken to the statutory mandate to eliminate systemic wage differentials between male and female employees performing work of equal value, it would not be an error of law for the Tribunal to construct a male comparator line, the average value of whose work exceeded the average value of the work performed by the complainant group.
[123] Fourth, to the extent that the Tribunal’s selection of the segmented line methodology involved the exercise of discretion, it clearly had a rational basis in the evidence. Expert witnesses testified that it was statistically sound, and that it was always preferable to adopt a methodology that incorporated more rather than less data.
[124] The statisticians were not troubled by the possible impact on the male regression line of the presence of “lonely observations” at the high-value end of the female regression line: Mr. Sunter, R.J.R., vol. VII, at pages 2391-2397; Dr. Shillington, loc. cit., at pages 2445-2446. Since the statisticians had already identified and discarded outlier observations that were statistically suspect, there was no reason to think that these observations were any less reliable than any other: Mr. Sunter, loc. cit., at pages 2311-2318, 2331-2336, 2348-2349; Dr. Shillington, loc. cit., at page 2502.
[125] Nor did they regard as statistically relevant the fact that the segmented line methodology included the “tails” of occupational groups, most members of which were performing work of greater value than that of anyone in the female complainant group: Dr. Shillington, loc. cit. , at pages 2400-2404, 2420-2421, 2484-2488.
[126] In contrast, the statisticians testified that it would be irrational for the Tribunal to select a methodology, such as the “whole group” methodology proposed by the Treasury Board, because it used too little of the available data: Dr. Shillington, loc. cit., at pages 2405-2407, 2421-2424, 2487, 2517-2519.
[127] Further, the data had not been collected during the JUMI study with a view to ensuring that it included a representative sample from each predominantly male occupational group in the federal public service: see, for example, exhibits PSAC-2 and PSAC-3, R.J.R., vol. V, at pages 1230-1233. Hence, it was not clear that the data were adequate for making statistically reliable findings about the relationship between the value of the work performed and the wages paid for whole occupational groups, the value of whose work had the same central tendency as that performed by employees in the complainant group.
[128] The evidence of the pay equity experts also supported the use of a methodology that used as much data as possible, which, of course, the whole group methodology does not: Dr. Weiner, R.J.R., vol. VI, at pages 2008-2009 and 2016. The Tribunal’s methodology was consistent with the methodologies used in other pay equity exercises with which they were familiar in jurisdictions where specialized pay equity legislation was in force.
(c) Conclusion
[129] For these reasons I can find nothing in the text of the Act, or in its underlying policy, that leads me to the conclusion that the Tribunal’s chosen methodology was inconsistent with the statute.
[130] The concept of systemic discrimination, the mischief at which section 11 is primarily aimed, can be difficult to grasp. As this case clearly shows, the elucidation and application of the principle of equal pay for work of equal value calls for the kind of multidisciplinary study in which the Tribunal engaged. I would also add that, as the exhibits referred to above indicate, until the breakdown of the JUMI, the Treasury Board had proceeded on the assumption that wage differentials between men and women would be measured on the basis of the composite line approach.
[131] While not free from its own technical difficulties, and backed by no expert evidence, the whole group methodology, advanced on behalf of the Attorney General as an example of a methodology that did comply with the Act, might nonetheless be adopted by the Tribunal in another case as appropriate for measuring a wage gap.
[132] Accordingly, on the basis of the evidence before it, I am satisfied that the methodology selected by the Tribunal, in the exercise of its discretion, for implementing the broad principle of equal pay for work of equal value in the employment context before it easily met the standard of rationality.
4. Causation
[133] Counsel for the Attorney General submitted that the methodology selected by the Tribunal was also legally flawed because it did not ensure that the only wage differences measured were those caused by sex discrimination. She emphasized that section 2 limited the purpose of the Act to the eradication of inequality caused by “discriminatory practices based on” [underlining added] the prohibited grounds. In this case, of course, sex is the relevant ground.
[134] Hence, the Tribunal was not authorized to redress wage differences between men and women that were the result of social conditions and structures that resulted in the underrepresentation of women in higher paying occupations. The Canadian Human Rights Act did not perform the broader functions of employment equity legislation.
[135] It followed from this, counsel submitted, that the inclusion in the Tribunal’s male comparators of men who were paid more than other men performing work of equal value invalidated the methodology, because these differences could not be based on sex. Accordingly, only the lowest paid of the predominantly male occupational groups performing work that had the same central tendency value as that of a complainant group could be used to measure any difference in wages between male and female employees performing work of equal value. Any greater wage difference between such employees is demonstrably not based on sex, but may be attributable to factors such as bargaining power or market demand.
[136] On the basis of this analysis of section 11 counsel argued that section 14 of the Equal Wages Guidelines, 1986 must be ultra vires because it provides that, when more than one occupational group is used for the purpose of comparison, the groups are deemed to be one. Thus, the Guidelines do not limit the permissible comparators to employees in the lowest paid predominantly male occupational group performing work of equal value to that of the complainant group.
[137] The Guidelines were issued pursuant to subsection 27(2) of the Act which authorizes the Commission to issue guidelines (which, I should add, are legally binding while in force: subsection 27(3)):
27. …
(2) The Commission may, on application or on its own initiative, by order, issue a guideline setting out the extent to which and the manner in which, in the opinion of the Commission, any provision of this Act applies in a class of cases described in the guideline.
[138] This is a very broad delegation of discretion which leaves to “the opinion of the Commission” the power to determine the manner and extent of the application of a provision of the statute to a described category of case. The provision in question here, of course, is section 11 which, as I have already indicated, from its terms and nature inevitably leaves considerable latitude in its implementation.
[139] I observe, too, that, rather unusually, Parliament has conferred on an independent administrative agency, the Canadian Human Rights Commission, the power to make subordinate legislation on substantive matters, as opposed to questions of practice and procedure. At the federal level such powers are more typically conferred on, or their exercise is subject to confirmation by, the Governor in Council.
[140] I infer from this broad delegation of authority that Parliament must have contemplated that the expertise that the Commission will have acquired in the discharge of its statutory responsibilities for human rights research and public education, and for processing complaints up to the point of adjudication, is more important than political accountability for ensuring the appropriate exercise of the legislative power conferred by subsection 27(2).
[141] In view of the breadth of the statutory language of subsection 27(2), and of the attributes of the body to which the discretion has been conferred, a provision in any guidelines issued will only be held to be invalid if it is clearly incompatible with the terms of the grant of statutory power, when construed in the light of the purposes of the Act, or is an unreasonable exercise of the Commission’s discretion. In my opinion, section 14 of the Guidelines is neither.
[142] The argument that section 14 is invalid can only succeed on the first ground if the purpose of the statute as set out in section 2 requires words to be read into section 11 of the Act. It will be recalled that subsection 11(1) merely prohibits an employer from maintaining or establishing “differences in wages between male and female employees employed in the same establishment who are performing work of equal value”.
[143] Counsel for the Attorney General argued that, since the definition of a discriminatory practice is one based on a prohibited ground, the discriminatory practice identified and proscribed by section 11, namely establishing or maintaining wage differences between male and female employees who are performing work of equal value, must by clear implication be limited to differences that are based on or caused by sex.
[144] If this interpretation were correct, then it would certainly be arguable that, for the Commission to issue guidelines that expand the definition of the conduct proscribed by subsection 11(1) would go well beyond legislating on the manner and extent of the application of a provision of the Act to a class of case. In my view, however, the premise of the argument is unsound: subsection 11(1) means what it says, and does not require words to be read into it.
[145] First, sections 5 to 10 [section 9 (as am. by S.C. 1998, c. 9, s. 12)] of the Act, which define other discriminatory practices, include as part of the definition that the proscribed conduct must be “based on a prohibited ground of discrimination”. Subsection 11(1), in contrast, does not state that the conduct prohibited, namely maintaining or establishing wage differences between male and female employees performing work of equal value, must be based on sex.
[146] It is arguable, of course, that, in a provision proscribing wage differences between male and female employees performing work of equal value, it would be redundant to add that the differences must be based on sex. This is obvious both from the subject-matter of section 11 and from section 2, where the purposes of the Act are set out. However, if a similar approach were taken to sections 5 to 10 it could also be said that, in light of section 2, it was unnecessary to include in the definition of the relevant conduct proscribed as a discriminatory practice that it was based on one of the prohibited grounds.
[147] Second, subsection 11(4) provides that, notwithstanding subsection 11(1), it is not a discriminatory practice to pay different wages to male and female employees performing work of equal value if that difference can be attributed to one of the factors prescribed in guidelines as a reasonable factor justifying the difference. Subsection 11(5) adds “[f]or greater certainty” that sex cannot be a reasonable factor justifying a wage difference.
[148] It can be inferred from subsection 11(4) that the definition of the particular discriminatory practice described in subsection 11(1) is not limited to differences in wages that are based on sex. If, as counsel for the Attorney General argued, section 11 only prohibits differences in wages paid to men and women performing work of equal value when that difference is on the basis of sex, subsection 11(4) would have been unnecessary.
[149] For example, paragraph 16(h) of the Guidelines exempts from section 11 of the Act differences in wages between men and women performing work of equal value that are attributable to the existence of an internal labour shortage in a particular job classification. If section 11 bears the narrow meaning that the Attorney General alleges, then a difference in wages caused by a particular labour shortage would have fallen outside the definition of a discriminatory practice because it would not have been caused by sex.
[150] That it was necessary to exempt from the reach of section 11 wage differences between men and women performing work of equal value that are attributable to prescribed “reasonable factors”, other than sex (subsection 11(5)), is a clear indication that they would otherwise have fallen within the scope of that section. Accordingly, once a complainant has established a difference in the wages paid to male and female employees performing work of equal value, a breach of section 11 is thereby established, subject only to the employer’s demonstrating that the difference is attributable to one of the “reasonable factors” prescribed in section 16 of the Guidelines.
[151] Third, the nature of systemic discrimination often makes it difficult to prove that the disadvantaged position in the workplace of many members of particular groups is based on the attributes associated with the groups to which they belong. This is because, as Dickson C.J. observed in Canadian National Railway Co., supra, at page 1139, systemic discrimination “results from the simple operation of established procedures of recruitment, hiring and promotion, none of which is necessarily designed to promote discrimination.” Accordingly, an employer’s wage policies and practices may be based on such deep-rooted social norms and assumptions about the value of the work performed by women that it would be extremely difficult to establish in a forensic setting that, if women were paid less than men performing work of equal value, that difference was based on sex.
[152] Subsection 11(1) can thus be seen to have tackled the problem of proof by enacting a presumption that, when men and women are paid different wages for work of equal value that difference is based on sex, unless it can be attributed to a factor identified by the Commission in a guideline as constituting a reasonable justification for it. By way of analogy it may be noted that, once a prima facie case of discrimination has been proved, the burden shifts to the employer to justify it by showing, for example, that it would impose undue hardship to accommodate the employee who had complained of discrimination: Ontario Human Rights Commission and O’Malley v. Simpsons-Sears Ltd. et al., [1985] 2 S.C.R. 536, at pages 558-559.
[153] Three other points may also be briefly noted here that detract from the causation argument made on behalf of the Attorney General. First, the Guidelines do not include “market forces” as a reasonable explanation of a wage differential, no doubt for the very good reason that the market is not itself free from discriminatory tendencies.
[154] Second, the lead management representative on the JUMI Committee, Ms. Lise Ouimet, a senior official at the Treasury Board, welcomed the “weighted average” approach when it was incorporated in section 14 of the Guidelines: R.J.R., vol. V, at page 1265. The Commission had adopted this approach in part because it had been used by the government when voluntarily settling pay equity claims in the late 1970s and early 1980s.
[155] Third, in order to establish unlawful discrimination on a ground prohibited by human rights legislation it is sufficient that the conduct in question was based in part on the prohibited ground. It does not have to be the only reason for the conduct.
[156] Accordingly, the structure and purpose of the Act do not require the addition to subsection 11(1) of words that are not there. Hence, section 14 of the Guidelines is not ultra vires the terms of the grant of power in subsection 27(2).
[157] Counsel for the Attorney General did not contend that, if section 14 were within the “four corners” of subsection 27(2) because it deals with “the manner and extent” of the application of subsection 11(1), it was invalid on the ground of unreasonableness. In my view, such an argument would have been untenable on the facts.
[158] I should also note in this context that provincial pay equity legislation does not adopt a uniform approach to this issue. On the one hand, the Ontario and Nova Scotia statutes expressly direct that the comparison is to be with the lowest paid male job class, while Manitoba, New Brunswick and Prince Edward Island require the adjustment to be to the average of the male wages.
[159] To conclude, section 14 of the Guidelines is authorized by subsection 27(2) of the Act and permits the Tribunal to do what it did here: that is, to treat the predominantly male occupational groups as one group for the purpose of measuring any wage differences between male and female employees performing work of equal value. It was not a breach of the Act for the Tribunal to have adjusted up the wages of the complainant group to the weighted average wage line of the male comparators.
5. “Occupational Groups”
(a) The Attorney General’s position
[160] Counsel for the Attorney General identified the Tribunal’s failure to comply with section 15 of the Equal Wages Guidelines, 1986 as the third fundamental flaw in its reasoning. She argued that this section requires that, where a complaint is made under section 11 of the Act by an “identifiable occupational group”, as happened here, any comparison of the value of work performed and wages paid must be to employees in other occupational groups. That is, the comparison must be to the occupational group as a whole, and not simply to employees who all happen to be in predominantly male occupational groups.
[161] The Tribunal’s methodology, she argued, did not take as the comparators of the employees in the complainant group employees in other occupational groups. The composite line of all employees in predominantly male occupational groups did not itself constitute an occupational group within the meaning of the Guidelines.
[162] Further, the segmented line could include fragments of occupational groups. This would occur when some employees of an occupational group were performing work that fell outside the value parameters of the complainant group, while others performed work that fell within them. The segmented line was thus not a comparison of occupational groups, but of employees who were performing work with a value that happened to fall within the range of values of that performed by members of the complainant group. This could not make them an “occupational group” for the purpose of the Guidelines.
[163] Further, counsel submitted, subsection 15(2) of the Guidelines provides that, for the purpose of making a comparison of the wages received by employees in the complainant occupational group, and employees of the occupational group or groups to which employees of the complainant group have compared themselves, the wage curve or regression line of employees in the comparator group or groups shall be used. Clearly, it was argued, by using a segmented line based on data collected on a position-by-position basis, the Tribunal had not done this, and had thereby committed an error of law.
[164] —Occupational groups” must refer to employee categories designated by the employer for wage and other employment and management purposes on the basis of the nature of the work performed by those employees. In this case, as I have already noted, the employer-created occupational groups within the federal public service were also used as the basis of the bargaining units established when collective bargaining was introduced.
[165] Thus, when the Alliance filed its complaint with the Commission in 1991 it did so on behalf of the Clerical and Regulatory occupational group, and compared the wages of the employees in that group to those paid to employees in 53 predominantly male occupational groups.
(b) The position of the Alliance and the Commission
[166] Counsel for the Alliance and the Commission submitted that, in rejecting the Treasury Board’s interpretation of section 15, the Tribunal had not committed an error of law. They argued that to use employer-created occupational groups to measure wage differences would not further the purpose of section 11 of the Act, namely, the elimination of differences in wages paid to male and female employees performing work of equal value as a result of systemic sex discrimination. Section 11 nowhere mentions “occupational groups” and they should therefore not be given the importance in the identification of wage differentials that was ascribed to them by the Treasury Board.
[167] Further, the occupational groups in the federal public service as designated unilaterally by the Treasury Board were quite unsuitable measuring units for pay equity purposes. For one thing, since they had been created more than 30 years ago there was a danger that they might be based on stereotypical assumptions about the relative values of work typically performed by men and women, and could thus be seen to be part of the problem of systemic wage discrimination, not part of the solution.
[168] In other respects, too, counsel submitted, the occupational groups were outmoded and were not even suited to perform the employment functions for which they were created. Thus, if by down-playing the significance of these occupational groups for wage comparison purposes the Tribunal’s methodology had the effect of disrupting “orderly salary-setting” within them, this was not necessarily to be regretted, because most such groups had long ago lost whatever coherence they at one time might have had as a basis for determining employee remuneration.
[169] For example, the definition of many occupational groups has not been amended in response to the changing nature of the work performed by employees assigned to them. Job descriptions often do not describe accurately the full range of duties actually performed by employees occupying particular positions within them. Some occupational groups, such as the Clerical and Regulatory complainant group, are very large and lump together employees performing very different tasks that involve work of widely varying values. Lastly, a significant minority of employees are assigned to an inappropriate occupational group.
(c) Analysis
[170] I have not found the Tribunal’s reasons on this issue easy to understand. The interpretation of section 15 of the Guidelines is considered in paragraphs 337-345 [pages D/467-D/468] of the Tribunal’s reasons for decision.
[171] The Tribunal’s discussion focusses almost exclusively on the fact that subsection 15(1) permits an indirect comparison of wages to be made in group complaints when a direct, job-to-job comparison is not possible, and that subsection 15(2) provides that the indirect comparison is to be made by means of the wage curve of the comparator occupational group.
[172] In paragraph 343 [page D/468] the Tribunal also says this:
We further find the words “occupational group” used in s. 15 of the Guidelines are intended in the same manner and for the same purpose as that which arises under s. 14 of the Guidelines and refers to groups designated by the application of s. 13 of the Guidelines as either female-dominated or male-dominated.
[173] The submissions of the Alliance and the Commission no doubt constitute a powerful case against the adoption of employer-designated occupational groups as the measuring units for comparing wage differences under section 11 of the Act. However, the question for me is whether, properly interpreted, the Guidelines prescribe that this is what must be done.
[174] Since the meaning of “occupational group” in the Guidelines is one of statutory interpretation I must determine on a standard of correctness whether the Tribunal erred in the conclusion that it reached. Further, because of the paucity of the Tribunal’s reasons on this issue I have found little there by way of elucidation. However, since the Commission promulgated the Guidelines, a responsibility entrusted to it by Parliament by virtue of its front-line experience with issues of discrimination, the submissions made on its behalf by counsel deserve particularly careful attention.
[175] The words “occupational group” recur throughout sections 12 to 15 of the Guidelines, sections that stipulate some aspects of the manner in which section 11 applies to complaints by groups. Although the words are not defined in the Guidelines, it is at least plausible to assume that they have the same meaning in each of these interlinked sections.
[176] It will be helpful at this point to recap some salient features of the statutory scheme and of the factual background to the present litigation. The Alliance’s complaint was made on behalf of an “identifiable occupational group” and compared the wages of employees in this group with the wages paid to employees in 53 other occupational groups in the federal public service. Both the complainant group and the comparator groups were occupational groups created by the employer.
[177] In accordance with sections 12 and 13 of the Guidelines a determination was then made that the complainant group was predominantly female, and that the 53 other occupational groups were predominantly male.
[178] The data available to the Commission, and subsequently to the Tribunal, about the wages paid to and the value of the work performed by employees in predominantly male occupational groups had been collected during the JUMI study. The data were representative of this population as a whole, because they were derived from questionnaires distributed to a random sample of employees selected by the positions that they occupied, and not by the occupational groups to which they belonged.
[179] Thus, the data enabled a statistically reliable picture to be constructed of the employer’s pay practices, or the relationship of wages to the value of work, for the totality of the employees working in predominantly male occupational groups. They did not, however, necessarily provide statistically reliable information about wages and work values for each occupational group or, consequently, for the 53 occupational groups cumulatively.
[180] Counsel for the Alliance and the Commission did not suggest a meaning that the words “occupational groups” bore in the Guidelines, other than occupational groups created by an employer. Rather, their position was that, because these groups were unsuitable for determining pay equity complaints, the Guidelines should be interpreted in a way that minimizes the role of occupational groups in the exercise. In particular, it was urged, occupational groups ceased to be legally relevant after they had been used to determine whether a complainant group was predominantly male or predominantly female in its composition and, if it was, to ensure that any comparator groups were predominantly of the other sex.
[181] Counsel for the Attorney General submitted that this was too narrow a view of the roles assigned by the Guidelines to occupational groups when complaints are made by groups. Thus, section 14 provides that, when a comparison is made with the work and wages of “other occupational groups , those other groups are deemed to be one group.” [Underlining added.]
[182] In addition, subsection 15(1) states that “[w]here a complaint alleging a difference in wages between an occupational group and any other occupational group is filed” and a direct comparison cannot be made between the value of the work performed and the wages “received by employees of the occupational groups”, an indirect comparison may be made of the work and wages “of each occupational group”. [Underlining added.]
[183] Most telling of all, perhaps, subsection 15(2) provides that, when an indirect comparison is made, “the wage curve of the other occupational group shall be used to establish the difference in wages, if any” between employees in the complainant “occupational group” and “the other occupational group”. [Underlining added.]
[184] Thus, counsel concluded, the Guidelines clearly contemplate that, for the purpose of determining under section 11 whether a wage difference exists, and if it does its size, the Tribunal must make its comparison between the complainant group and employees in occupational groups that are predominantly of the other sex. In order to perform this exercise the Tribunal must have data that enable it to compare on the basis of occupational groups. The data available to the Tribunal in this case did not enable it to make this kind of comparison.
[185] Despite what counsel for the Attorney General submitted was the “plain meaning” of these provisions, if the Guidelines are interpreted in light of the Tribunal’s interpretation of section 11 of the Act, which I have upheld, it is difficult to see what purpose would be served by a requirement in the Guidelines that the comparison be made on the basis of occupational group, as opposed to one based on the employees of predominantly male occupational groups, sampled by position and not differentiated by occupational group.
[186] In any event, given the large number of predominantly male occupational groups involved here, it is unlikely that the regression line drawn to summarize the wages and value of work data of the employees sampled by occupational group, melded into a composite line pursuant to section 14 of the Guidelines and trimmed to a segment, would differ much from the line produced from data derived from a random sample of the employees by position.
[187] Accordingly, in my opinion the references to “occupational groups” in the sections of the Guidelines simply refer to the groups identified under sections 12 and 13 as predominantly male or predominantly female. Interpreted both in the context of section 11 properly understood, and of the dangers inherent in assigning to employer-designated occupational groups any broader significance in the pay equity exercise, I cannot conclude that section 15 mandates that comparisons be based on employees in predominantly male occupational groups, sampled by group.
[188] In the event that I am wrong on this point, and the Tribunal did commit an error of law because section 15 of the Guidelines requires the Tribunal to base its conclusion on the wage curve of the predominantly male occupational groups, the error would not warrant the quashing of the Tribunal’s decision. I discuss this issue later under the heading, “G. REMEDIAL DISCRETION”.
6. Wage Ratcheting
[189] An argument advanced before the Tribunal by the Treasury Board, and repeated in this proceeding by counsel for the Attorney General, was that the methodology proposed by the Commission, and eventually adopted by the Tribunal, would inevitably lead to the ratcheting of wages. The result would be that one pay equity adjustment would inevitably lead to another, and that to yet another, and so on, until all employees were eventually paid wages equal to those of the most highly paid employees performing work of equal value, regardless of sex.
[190] This would occur because the Tribunal’s methodology requires the wages of the female occupational group to be moved up to the regression line representing the average wages paid to employees in predominantly male occupational groups. By definition, there will be male employees whose wages are below the male average, and consequently below the wages paid to females following a section 11 wage adjustment.
[191] Section 11 may be invoked by male and female employees alike. Accordingly, males whose wages are below the male average and now, therefore, following a section 11 wage adjustment, below those of females performing work of equal value, may complain that there is a wage difference between male and female employees performing work of equal value.
[192] Moving up these males to the average male wage line would, of course, cause the male average wage line to rise. This would set up a difference in the wages paid to male and female employees performing work of equal value because the wages paid to females would now be below the new male average. This would support another complaint by females and result in an adjustment of their wages up to the new male average. The males below the newly-adjusted female wage would then make a further section 11 complaint and their wages would be adjusted to the new male average. The process would end only when all employees, regardless of sex, were paid the same top rate for work of equal value.
[193] Counsel for the Attorney General submitted that this issue was specifically addressed in Syndicat des employés de production du Québec et de l’Acadie v. Canada (Canadian Human Rights Commission), [1989] 2 S.C.R. 879 (SEPQA), where Sopinka J. endorsed a particular interpretation of section 11 because it would avoid wage ratcheting which, he said (at page 888), was contrary to the purpose of the Act because:
…. s. 11 would not be a guarantee of equal pay between sexes, but a guarantee of equal pay for work of equal value irrespective of sex.
[194] Counsel submitted that ratcheting could not occur under the methodology proposed by the government for eliminating wage differences between men and women because the wages of a predominantly female complainant occupational group would only be raised to the level of the lowest paid predominantly male occupational group performing work of equal value.
[195] The Tribunal rejected this argument and stated at paragraph 393 [page D/491]:
The rationale behind a wage adjustment methodology in the complaints before the Tribunal is the establishment of a male wage curve to capture the male wage policy of the employer. This cannot, in our view, give rise to a claim by males of a discriminatory practice. Ratcheting, in this instance, does not capture the essence of what pay equity is designed to achieve and in that sense is not compatible with the purpose of s. 11.
[196] In my opinion this is eminently good sense. However, it cannot be determinative of the issue before me that ratcheting appears not to have been a problem in the context of other pay equity legislation where a methodology has been used for measuring differences similar to that employed by the Tribunal. Such statutes typically permit only predominantly female groups to complain.
[197] Nor, with respect, is it sufficient for counsel for the Commission to assert that it would not accept a section 11 complaint by male employees when the difference in wages alleged between males and females resulted from a previous section 11 adjustment following a complaint by females.
[198] The relevant question is whether a claim by male employees in the hypothetical under consideration here would amount to an allegation that the employer was establishing or maintaining a difference in wages paid to male and female employees performing work of equal value contrary to section 11 as properly interpreted.
[199] In my opinion it would not. While section 11 certainly does not preclude complaints by male employees that they are paid less than female employees performing work of equal value, no interpretation of section 11 can ignore the fact that the mischief at which it is principally aimed is the existence of a wage gap that disadvantages women, as a result of gendered segregation in employment and the systemic undervaluation of the work typically performed by women.
[200] Moreover, section 14 of the Guidelines is an enactment of the Commission’s view that the principle of equal pay for work of equal value does not require the adjustment of female wages up to the level of the highest paid males performing work of equal value. It is sufficient if female wages do not fall below the average of those received by male employees.
[201] Accordingly, a complaint by some of the male employees that they are paid less than females whose wages have been adjusted up to the average male wage pursuant to section 11, and that they are thus the victims of a difference in wages contrary to section 11, could not be upheld consistently with the statutory scheme. A literal interpretation of the words of section 11 must yield to an interpretation that is informed by the statutory context in which they appear.
[202] When the wages paid to female employees are adjusted upwards pursuant to a section 11 complaint, and in accordance with the methodology used by the Tribunal in this case, any wage difference thereby created between the wages of females and the wages of males whose wages are below the average could appropriately be regarded as statutorily authorized. It is accordingly not a difference that the employer has established or maintained within the meaning of subsection 11(1).
[203] It could also be said that, when a section 11 wage adjustment is made in response to a group complaint, any sex-based wage lines disappear. The result is to produce a single line for all groups regardless of whether they are predominantly male or predominantly female. In such a situation, when a systemic remedy has been fashioned there is no room for an individual complaint under section 11.
[204] Nor is this conclusion inconsistent with SEPQA, supra, where the facts were very different from those before me, and the principal issue concerned the procedural obligations of the Commission and whether its decisions were reviewable under section 18 or section 28 of the Federal Court Act [R.S.C. 1970 (2nd Supp.), c. 10] prior to its amendment in 1993.
[205] Further, it would seem that, in the passage where the sentence quoted above appears, Sopinka J. was merely explaining, without necessarily adopting, the manner in which the Commission had avoided ratcheting. The interpretation of section 11 that I have adopted also avoids ratcheting, although not in the manner apparently referred to in SEPQA, supra.
F. RECALCULATION
[206] It was common ground at the hearing of the application for judicial review that, if the data collected in the JUMI study revealed that a wage gap existed in 1987 in contravention of section 11 of the Act, any monetary award made to close that gap should be retroactive from the date of the Tribunal’s decision, July 29, 1998, to March 1985 when the JUMI study was announced and the respondent agreed to hold in abeyance existing and potential pay equity complaints made to the Commission. The parties disagreed, however, about how the amounts payable in each of those years, other than 1987, should be calculated.
[207] The Tribunal adopted the methodology proposed by the Commission and the Alliance. This involves a recalculation of the wage gap for each year by taking into account salary increases paid not only to members of the complainant group, but also to employees in the predominantly male occupational groups included in the segmented line. Thus, the wage gap would increase if the salaries of the comparator employees rose faster than those of the complainants, and shrink if the complainants’ pay increases were larger.
[208] The Attorney General seemed not to object in principle to this method of calculation, but counsel argued that it could not lawfully be used on the facts of this case. This is because it assumes that the value of the work performed by those whose wages are being compared remained constant throughout the period of retroactivity. However, evidence was given by Mr. Willis that, as a result of the changing nature of work, the values of work performed over, for example, a 10-year period typically would not remain constant. Thus, the further one moves from 1987, the year to which the JUMI study data related, the less reliable the data would be for determining the value of the work performed by employees in the complainant and comparator occupational groups.
[209] Hence, to measure the wage gap in any given year by reference to, say, 1995 salary levels and 1987 values of the work performed would not be in accordance with section 11 of the Act because the measurement made of the wage gap would not have taken into consideration changes likely to have occurred in the value of the work performed. Accordingly, the wage difference measured by the Tribunal’s method of calculation might not constitute a difference in wages between male and female employees performing work of equal value. Any wage differential in these years could thus not be attributed to gender.
[210] Counsel therefore submitted that the wage gap should be calculated once and for all in 1987, and increased by any percentage increase in the salaries paid to members of the complainant group. Thus, on the Treasury Board’s formula, if a 2 per cent increase were awarded to members of a complainant group in 1988 as a result of collective bargaining, the sum payable for that year would be the 1987 wage gap plus 2 per cent, regardless of what happened to the salaries of the comparators from the predominantly male occupational groups.
[211] The weakness of this proposal, of course, is that it does not purport to do what the Act requires, which is to measure differences in wages between male and female employees performing work of equal value.
[212] It was conceded that, while wage rate data were available, it would be impossible to determine the actual value of the work performed by members of the complainant and comparator groups in the years after 1987. The absence of relevant data undoubtedly created a serious practical problem for the Tribunal. However, the Tribunal had to do the best that it could with the data that it had to implement the provisions of section 11, interpreted in a broad and liberal manner.
[213] In my opinion, it was open to the Tribunal in these circumstances to adopt the annual recalculation method for determining the wage gap in each of the years covered by the retroactive award period. In reaching this conclusion I have also taken into account the fact that the Treasury Board’s proposed method of calculating the wage gap proceeds on the equally unlikely assumption that the wage gap remained constant during that period.
[214] In support of the method chosen by the Tribunal it may be pointed out that the Treasury Board itself had agreed to the adoption of the annual recalculation method in some pay equity settlements that it had made with the Alliance. In addition, when the Treasury Board made a unilateral pay equity award in 1990 it also calculated wage gaps by the annual recalculation method subsequently adopted by the Tribunal in this case.
[215] Indirect support for the Tribunal can also be found in Public Service Alliance of Canada v. Canada (Department of National Defence), [1996] 3 F.C. 789 (C.A.), where Hugessen J.A. stated that it was consistent with the principles of human rights legislation to presume that, once a wage gap was shown to exist, it had also existed prior to the period studied. Consequently, once the Alliance had demonstrated that, on the basis of 1987 data, a wage differential of a particular size existed, the employer bore the burden of showing that changes had occurred in the value of the work performed or wages paid that would have altered the size of the wage gap for the purpose of section 11.
[216] The evidence of Mr. Willis that changes in the type of work performed over a period of more than 10 years were likely to have diminished the reliability of the 1987 data as it related to the value of the work performed by the employees being compared is in my opinion insufficient to discharge the Treasury Board’s burden of proof and thereby to render the Tribunal’s decision an unreasonable way of implementing the broad purposes of section 11. There was no evidence whether any such changes had in fact occurred and, if they had, whether they widened or narrowed the wage differential.
[217] The following observation of Hugessen J.A. (at page 815) is equally apposite to the position in which the Tribunal found itself in this case:
… once it is known that a plaintiff has suffered damage, a court cannot refuse to make an award simply because the proof of the precise amount thereof is difficult or impossible. The judge must do the best he can with what he has. [Emphasis added.]
[218] In summary, in light of the imperfect information available to it, including an absence of any evidence adduced by the Treasury Board that changes in the value of work performed by the employees in question had in fact occurred, I cannot characterize as unreasonable the methodology adopted by the Tribunal, especially since the Treasury Board’s alternative came no closer to a faithful implementation of section 11.
[219] For these reasons, it is my opinion that the Tribunal committed no error of law in adopting the annual recalculation method for determining the size of the wage gap in the period 1985 to 1998.
G. REMEDIAL DISCRETION
[220] If, contrary to the view that I expressed earlier, the Tribunal did err in law by failing to compare the wages paid and the value of the work performed by employees in the complainant occupational group with those of predominantly male occupational groups in the manner directed by the Guidelines, it does not necessarily follow that the Tribunal’s decision must be set aside.
[221] On an application for judicial review it is within the Court’s discretion to grant or to refuse the relief sought by an applicant, even when a reviewable error has been committed by the administrative decision-maker. There are a number of grounds on which a remedy may be withheld in the exercise of the Court’s discretion, some of which are potentially relevant here.
[222] Thus, misconduct by the applicant may warrant the refusal of relief. Counsel for the Alliance strenuously argued that this was the case here.
[223] In particular, he relied on the fact that, during the JUMI study the Treasury Board had proceeded on the assumption that the male comparator line would be constituted as a composite of all employees in predominantly male occupational groups, sampled by the work that they performed, not by their occupational group. The Treasury Board could not therefore subsequently allege before the Tribunal and this Court that the male comparator line had to be constructed on the basis of occupational group.
[224] However, in my view this is not the kind of conduct that precludes the government, as the representative of the public interest, from subsequently asserting in good conscience that, on a closer examination of the legislation, it was now taking a different view of what the law required. As counsel for the Attorney General fairly put it, during the JUMI study all parties were learning about the intricacies of identifying and measuring systemic wage discrimination, their agreement on methodology was no more than tentative, there was no jurisprudence interpreting the relevant statutory provision, and the Alliance and the Treasury Board at that time were working in a co-operative mode.
[225] Nor am I particularly impressed by the evidence indicating that the government only changed its mind about the restrictions on the methodology imposed by the Act and the Guidelines when it realized that the data on the value of the work performed by employees in the female complainant group indicated that the wage difference was much greater than it had previously believed. A concern on the part of the government not to spend more public money than the law requires, or authorizes, is quite legitimate.
[226] For the same reason I do not think that it was misconduct for the Treasury Board to take issue before the Tribunal with the reliability of the evaluations of the work performed by employees in the complainant groups, or with the use by the Tribunal of data collected in response to confidential questionnaires. Although ultimately unsuccessful, neither challenge was frivolous.
[227] It was also argued that the government was now, in effect, estopped from challenging the validity of the Tribunal’s decision as a result of a statement made in 1993 prior to the election by Mr. Chrétien when Leader of the Opposition. He had said that if a Liberal Government were elected it would implement without further delay any decision of the Tribunal in this dispute. Whether or not, as counsel contended, this statement should be regarded as intended to apply only to a Tribunal decision that was lawful, accountability for such statements is in my view more appropriately enforced through the political process than through the courts.
[228] The most powerful reason for withholding relief is that, to set aside the Tribunal’s decision on the ground that it did not comply with section 15 of the Guidelines, would likely frustrate the purposes underlying section 11 of the Canadian Human Rights Act, and inflict a substantial injustice on the thousands of employees in the federal public service, many of whom are now retired, who would be deprived of the payment of back-wages, future wages and pension increases to which they are entitled.
[229] To be weighed against these considerations, of course, is the principle that administrative tribunals are obliged to exercise their statutory decision-making powers in accordance with their enabling legislation, and that public money should not be disbursed pursuant to decisions that are inconsistent with the legally binding instructions of the legislature.
[230] In this context it is important to remember that an application for judicial review is a public law proceeding and that in the final analysis relief is granted by the court in order to further the public interest. Thus, relief should be refused when it would not serve the public interest to set aside a decision, even if vitiated by legal error. This is the basic principle that informs the various grounds on which the discretionary remedies available on an application for judicial review may be withheld.
[231] In striking the balance between the competing heads of public interest outlined above, I have taken into account the following considerations.
[232] First, for all practical purposes it would be impossible for the comparator population to be re-sampled by occupational group. Detailed questionnaires would have to be administered in an attempt to discover the value of the work that employees in these groups were performing at the relevant time, that is more than ten years ago. In view of this substantial lapse of time, which would also have accounted for retirements and departures from the public service for other reasons by employees in the relevant occupational groups, it is highly improbable that reliable data could be assembled from which the value of the work performed by these groups could be determined.
[233] Second, even if theoretically possible, re-sampling the male comparators by occupational group would involve considerable expense and entail further delay in the resolution of this dispute. This is a matter that has also dragged on for far too long, and at far too great a cost for all concerned. I would be reluctant to grant a remedy that would have the effect of imposing further delay, with the consequent injustice that this would inflict on many. Justice unduly delayed in this context is indeed likely to be justice denied.
[234] As for the public interest in legality and the lawful expenditure of public money, I cannot say that, if the Tribunal erred in law in failing to comply with the Guidelines, the error was immaterial; there was no evidence on what difference it would have made to the ultimate outcome. However, I would not expect any difference be large, but I cannot be sure. Nor, on the other hand, do I know which party it would favour.
[235] Finally, for the reasons that I have already given, if the Tribunal did breach section 15 of the Guidelines, it was an error of a merely technical nature that did not thwart the essential purposes of section 11 or their implementation.
[236] Balancing these competing considerations, I have concluded that the benefit to the public interest of setting aside the Tribunal’s decision for failure to comply with section 15 would be outweighed by the costs of so doing, even if I were to remit the matter for the sole purpose of requiring the correction of this possible error.
H. CONCLUSION
[237] In my opinion the position taken by the Attorney General in these proceedings contains two structural flaws. First, its approach to the interpretation of the Canadian Human Rights Act and the Equal Wages Guidelines, 1986 is too abstract: it is insufficiently grounded in the factual realities of the employment context under consideration, the testimony of the array of expert witnesses who assisted the Commission and Tribunal, or analogous legislation in other jurisdictions.
[238] The Attorney General has sought to convert into questions of general law and statutory interpretation aspects of the implementation of Parliament’s enactment of the principle of equal pay for work of equal value that are better regarded as factual, technical or discretionary issues, or questions of mixed fact and law, entrusted to the specialist agencies responsible for administering the legislation.
[239] Second, the Attorney General’s argument was based on the narrowest possible interpretation of the Canadian Human Rights Act, including the definition of the problem at which section 11 was aimed and the measures that the Tribunal could lawfully take to tackle it. It paid only lip service to the regular admonitions from the Supreme Court of Canada that, as quasi-constitutional legislation, human rights statutes are to be interpreted in a broad and liberal manner.
[240] The Attorney General too often seemed to regard the relevant provisions of the Act as a straitjacket confining the Tribunal, instead of as an instrument for facilitating specialist agencies’ solution of long-standing problems of systemic wage differentials arising from occupational segregation by gender and the undervaluation of women’s work.
[241] Three consequences follow from the Attorney General’s interpretation of the Canadian Human Rights Act and the Guidelines. First, the Tribunal is prohibited from using a methodology for identifying and measuring a wage differential that would take into account the fact that women are underrepresented among employees who are performing more highly valued work, for which the remuneration increases more rapidly than the value of the work. Instead of a methodology approved by all the expert witnesses from whom the Tribunal heard, it should have adopted one for which there was no support at all from any witness.
[242] Second, the Tribunal is prohibited from identifying and measuring a wage differential by comparing the wages paid to employees in predominantly female occupational groups with the average wage paid to employees in predominantly male groups who are performing work of equal value. Instead, the Tribunal must confine its comparison to the wages of employees in the lowest paid male group.
[243] Third, the Tribunal is required to base its comparison on occupational groups, despite their limited utility as a basis for setting salaries in general and for pay equity exercises in particular.
[244] In my opinion Parliament cannot be taken to have required any of these consequences when it enacted the principle of equal pay for work of equal value in section 11 of the Canadian Human Rights Act in an attempt to eradicate systemic wage discrimination resulting from the gendered segregation of work and the undervaluation of the work typically performed by women.
[245] For the reasons that I have given, the application for judicial review is dismissed.
[246] Largely on the basis of the success of the Alliance in this proceeding I award it its costs, despite the importance of the issues raised by the applicant. However, I do not accept the Alliance’s submissions that costs on the highest scale are warranted.